BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:The Human Computer DTSTART;VALUE=DATE-TIME:19761210T150000Z DTEND;VALUE=DATE-TIME:19761210T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-178204@cern.ch DESCRIPTION:http://indico.cern.ch/conferenceDisplay.py?confId=178204 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=178204 END:VEVENT BEGIN:VEVENT SUMMARY:Campeonato de Mus DTSTART;VALUE=DATE-TIME:19900803T060000Z DTEND;VALUE=DATE-TIME:19900803T070000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-19727@cern.ch DESCRIPTION:http://indico.cern.ch/conferenceDisplay.py?confId=19727 LOCATION:CERN 1-1-025 URL:http://indico.cern.ch/conferenceDisplay.py?confId=19727 END:VEVENT BEGIN:VEVENT SUMMARY:T2 - marked up 1 DTSTART;VALUE=DATE-TIME:19900804T060000Z DTEND;VALUE=DATE-TIME:19900804T070000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-19728@cern.ch DESCRIPTION:\nHPC has undergone a revolution with the advent of commodity compute\nclusters. Where once HPC was the domain of a specialized few rese archers\,\nnow virtually any enterprise can gain access to large aggregate s of\nprocessing power - holding out the promise of truly "personal\nsuper computing."Personal supercomputing will enable scientists and\nengineers t o interactively employ massive computational and data processing\nresource s to solve complex technical problems.\n\n\nThey will be able to assemble\ ncomplex models from best of breed scientific applications and most recent \ndata sources that might span organizations and institutions\, while\nint egrating with the collaborative and business workflows.\n\n\nhttp://indico .cern.ch/conferenceDisplay.py?confId=19728 LOCATION:CERN 1-1-025 URL:http://indico.cern.ch/conferenceDisplay.py?confId=19728 END:VEVENT BEGIN:VEVENT SUMMARY:T3 - marked-up 2 DTSTART;VALUE=DATE-TIME:19900805T060000Z DTEND;VALUE=DATE-TIME:19900805T070000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-19729@cern.ch DESCRIPTION:\nHPC has undergone a revolution with the advent of commodity compute\nclusters. Where once HPC was the domain of a specialized few rese archers\,\nnow virtually any enterprise can gain access to large aggregate s of\nprocessing power - holding out the promise of truly "personal\nsuper computing."Personal supercomputing will enable scientists and\nengineers t o interactively employ massive computational and data processing\nresource s to solve complex technical problems.\n\n\nThey will be able to assemble\ ncomplex models from best of breed scientific applications and most recent \ndata sources that might span organizations and institutions\, while\nint egrating with the collaborative and business workflows.\n\n\nhttp://indico .cern.ch/conferenceDisplay.py?confId=19729 LOCATION:CERN 1-1-025 URL:http://indico.cern.ch/conferenceDisplay.py?confId=19729 END:VEVENT BEGIN:VEVENT SUMMARY:T4 - marked up with CSS DTSTART;VALUE=DATE-TIME:19900806T060000Z DTEND;VALUE=DATE-TIME:19900806T070000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-19732@cern.ch DESCRIPTION:Dear Miguel Angel Marquina\,\n \nYour request has been registered into the IT Problem Tracking System with\nthe ticket number: CT447464. Fo r all future communication related to this\nquery select link\,or use the reply address\, quoting ticket number CT447464). You will receive\nfurther information once the responsible person for the relevant support\nline ex amines your request.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId= 19732 LOCATION:CERN 1-1-025 URL:http://indico.cern.ch/conferenceDisplay.py?confId=19732 END:VEVENT BEGIN:VEVENT SUMMARY:CORBA Tutorial DTSTART;VALUE=DATE-TIME:19990608T120000Z DTEND;VALUE=DATE-TIME:19990608T123000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033095@cern.ch DESCRIPTION:This tutorial aims to give a practical introduction to the Obj ect Management Group's (OMG) Common Object Request Broker Architecture (CO RBA). \n\nThe Common Object Request Broker Architecture (CORBA)\, is the O bject Management Group's answer to the need for interoperability among the rapidly growing number of hardware and software products available today. With CORBA\, users gain access to information transparently\, without hav ing to know on what software or hardware platform it resides or where it i s located on enterprises' network. Simply stated\, CORBA allows applicatio ns to communicate with one another no matter where they are located or who has designed them.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a 033095 LOCATION:CERN 40-SS-A01 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033095 END:VEVENT BEGIN:VEVENT SUMMARY:Simulation of Distributed Processing Systems DTSTART;VALUE=DATE-TIME:19990609T140000Z DTEND;VALUE=DATE-TIME:19990609T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033096@cern.ch DESCRIPTION:The computing models for the LHC experiments envisage a comple x set of data servers connected over wide area\, regional and local area n etworks\, creating a distributed architecture computing systems. Systems o f even similar complexity will not exist for several years to come\, and a joint R&D project (MONARC) was started last year to develop the simulatio n tools and to perform feasibility studies of distributed architecture sys tems for LHC experiments analysis and computing. A simulation program\, bu ilt using Java technology\, which provides adequate tools for developing a flexible and powerful distributed process oriented simulation\, is descri bed.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033096 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033096 END:VEVENT BEGIN:VEVENT SUMMARY:Introduction to the UML (1/2) DTSTART;VALUE=DATE-TIME:19990830T070000Z DTEND;VALUE=DATE-TIME:19990830T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-226203@cern.ch DESCRIPTION:http://indico.cern.ch/conferenceDisplay.py?confId=226203 LOCATION:CERN Salle Dirac URL:http://indico.cern.ch/conferenceDisplay.py?confId=226203 END:VEVENT BEGIN:VEVENT SUMMARY:The Strengths\, Weaknesses\, Opportunities\, and Threads of Object Technology (2/2) DTSTART;VALUE=DATE-TIME:19990901T120000Z DTEND;VALUE=DATE-TIME:19990901T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-226204@cern.ch DESCRIPTION:http://indico.cern.ch/conferenceDisplay.py?confId=226204 LOCATION:CERN Salle Anderson URL:http://indico.cern.ch/conferenceDisplay.py?confId=226204 END:VEVENT BEGIN:VEVENT SUMMARY:The Workbook Approach: Getting New Users Up to Speed in a HEP Offl ine Environment. DTSTART;VALUE=DATE-TIME:19990901T140000Z DTEND;VALUE=DATE-TIME:19990901T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033097@cern.ch DESCRIPTION:The Workbook approach is an efficient way to introduce members of a HEP collaboration to their collaboration's offline environment. It was first developed in SLD where offline experts were losing too much time on individual user assistance.\nThe approach was extremely successful the re and has since been replicated for BaBar.\nInformation is presented in t he form of a workbook\, a connected series of exercises. Imagine the world of offline software as a workshop full of tools.\nThe workbook has the us er pick up each tool in turn\, do something simple and useful with that to ol\, and then put that tool down and move on to the next.\nReferences:\nTh e Workbook for BaBar Offline Users:\nhttp://www.slac.stanford.edu/BFROOT/w ww/doc/workbook/workbook.html\nThe Workbook for SLD Offline Users:\nhttp:/ /www-sld.slac.stanford.edu/sldwww/workbook/workbook_prod.htmlOrganiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a03309 7 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033097 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS TECHNICAL SESSION:
CAST Workbench - un outil de cartographie d’application DTSTART;VALUE=DATE-TIME:19990902T080000Z DTEND;VALUE=DATE-TIME:19990902T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033098@cern.ch DESCRIPTION:– Présentation de la société CAST et positionnement du pro duit\n– Présentation des fonctionnalités principales de CAST Workbench sur une application\n Compréhension de l’application par navigation entr e les liens Objet-Objet (SP\, function\, tables\, vues\, triggers\, séque nces\, synonymes\, écrans clients\, etc.)\n Analyse d’impact sur toute l ’application du changement d’un objet\n Développement de code\, debug\, specification\, standards de développements\, suivi d’avancée de proje t \n– Question Réponse techniques Organiser(s): Pedro Martel/EST-ISS\n\n http://indico.cern.ch/conferenceDisplay.py?confId=a033098 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033098 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS TECHNICAL SESSION:
QUEST Software Products\n DTSTART;VALUE=DATE-TIME:19990920T120000Z DTEND;VALUE=DATE-TIME:19990920T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033099@cern.ch DESCRIPTION:- Overview Quest's current product palette.\n\n- SQLAB Expert presentation.\nPL/SQL Application tunning: why is response time poor? It w ill point the exact SQL commands using most of the system's resources\, sh ow the best alternatives\, and create appropriate SQL itself.\n\n- SQLLAB Expert interface to SQL IMPACT.\n\n- SQL IMPACT presentation.\nYear 2000 i mpact analysis with automatic creation of E-R diagrams. For a change in a table\, IMPACT will tell which programs/procedures are affected\, and vice -versa. Also Year 2000 'detect and correct' functionality to debug mileniu m problems.\n\n- QuestionsOrganiser(s): Pedro Martel/EST-ISSNote: * Please note the unusual hours\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?co nfId=a033099 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033099 END:VEVENT BEGIN:VEVENT SUMMARY:High-performance Data Intensive Distributed Computing DTSTART;VALUE=DATE-TIME:19990923T083000Z DTEND;VALUE=DATE-TIME:19990923T120000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033100@cern.ch DESCRIPTION:The Data Intensive Distributed Computing Group at LBNL has dev eloped some distributing computing technologies that we feel could be extr emely useful to the HENP community. \n\nThe first is the Distributed Paral lel Storage System (DPSS)\, which provides cost effective parallel access to many disks\, and is tuned for wide-area network access. \nWe have demon strated a HENP data analysis application with read and write speeds of up to 57 MBytes/sec over a OC-12 WAN using a 4 server DPSS system. Using a DP SS\, an application can access remote data even faster than it can access data from a local disk\, eliminating the need to copy data sets from the r emote site before processing or visualising it. \n\nFor more information\, see: \nhttp://www-didc.lbl.gov/DPSS/\n\nThe second is The NetLogger Toolk it\, which includes tools for instrumenting applications\, tools for monit oring hosts and networks\, and tools for collecting and visualising monito ring data. Using these tools enables us to do end-to-end monitoring\, and detailed performance analysis of every component in the system. \nThis typ e of analysis is critical for obtaining performance in widely distributed systems. \n\nFor more information\, see: \nhttp://www-didc.lbl.gov/NetLogg er/\nOrganiser(s): G.FOLGER / IT\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=a033100 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033100 END:VEVENT BEGIN:VEVENT SUMMARY:Java-based Linear Collider Detector Simulations DTSTART;VALUE=DATE-TIME:19991020T140000Z DTEND;VALUE=DATE-TIME:19991020T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033101@cern.ch DESCRIPTION:The Java Analysis Studio (JAS\, pronounced jazz...) and the he p.lcd class library provide a general framework for performing Java-based distributed analysis. The package is being developed to fully reconstruct 500 GeV to 1.5 TeV annihilation events for studying the e+e- and m+m- dete ctor options for future colliders. The current U.S. effort includes first- pass pattern recognition\, vertex finding\, clustering\, track-cluster ass ociation and jet finding for use in detailed detector studies. This talk w ill describe the JAS/hep.lcd distributed analysis framework\, some aspects of the reconstruction object modeling\, a concrete example of our trackin g code\, and \nshow our latest effort to reconstruct W's from hadronic jet s using various energy flow algorithms.\n\nOrganiser(s): G.Folger / IT\n\n http://indico.cern.ch/conferenceDisplay.py?confId=a033101 LOCATION:CERN URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033101 END:VEVENT BEGIN:VEVENT SUMMARY:The State-of-the-Art in Magnetic Tape Data Recording\n DTSTART;VALUE=DATE-TIME:19991021T120000Z DTEND;VALUE=DATE-TIME:19991021T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033102@cern.ch DESCRIPTION:Digital magnetic tape remains the lowest cost method of storin g data and information for computing. A brief history of magnetic tape dat a recording will be presented.\n\nCurrent state-of-the-art magnetic tape s ystems will be reviewed\, with particular emphasis on fast-access linear d esigns. Trade-offs with the supporting technologies of recording channels\ , tape paths\, recording and reproduce heads\, servo control systems\, and the media itself will be discussed. Future developments hold the promise of still lower costs.\n\nAbout the speaker\n\nCharlie Farel is vice presid ent of Enterprise Nearline Development for Storage Technology Corporation. He is responsible for developing new tape drives and automated tape libra ry products and solutions for StorageTek's enterprise customers.\n\nSince joining StorageTek in 1981\, Charlie has held engineering management posit ions in optical disk\, magnetic disk\, tape drives and libraries\, and adv anced technology. \n\nCharlie holds degrees in electrical engineering from the University of Minnesota and Stanford University.\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=a033102 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033102 END:VEVENT BEGIN:VEVENT SUMMARY:C++ Code Analysis: an Open Architecture for the Verification of Co ding Rules\n DTSTART;VALUE=DATE-TIME:19991125T100000Z DTEND;VALUE=DATE-TIME:19991125T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033103@cern.ch DESCRIPTION:The analysis of C++ code is the basic building block of the co llaboration between ITC-IRST and CERN/ALICE\, aimed at improving the quali ty of the software by exploiting the information that can be automatically gathered from the code.\n\nThe first objective of the collaboration is th e development of a coding rule check tool. Successive steps will include a reverse engineering module and an intelligent refactoring tool. Since all \nplanned applications\, and possibly also those not yet considered\, sha re a common analysis bulk\, particular attention was devoted to the develo pment of an open architecture for the analysis of C++ code. In this talk t he adopted architectural solutions are presented and discussed\, highlight ing their generality\, the possibilities of extension that they offer\, an d the way details could be encapsulated within packages\, so that a clear and sharp interface between the subsystems is defined. The peculiarities o f the C++ language are also described\, together with the way they were ap proached and the state of the current implementation.\n\nShort biography\n \nPaolo Tonella received his laurea degree cum laude in Electronic Enginee ring from the University of Padua\, Italy\, in 1992\, and his PhD degree i n Software Engineering from the same University\, in 1999\, with the thesi s ÒCode Analysis in Support to Software MaintenanceÓ. Since 1994 he has been a full time researcher of the Software Engineering group at IRST (Ins titute for Scientific and Technological Research)\, Trento\, Italy. He par ticipated in several industrial and European \nCommunity projects on softw are analysis and testing. His current research interests include software engineering\, reverse engineering\, object oriented programming and code a nalysis.Organiser(s): F.Gagliardi / ITNote: Please note unusual DAY and TI ME\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033103 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033103 END:VEVENT BEGIN:VEVENT SUMMARY:The use of Globus for heterogeneous and distributed applications\n DTSTART;VALUE=DATE-TIME:19991201T150000Z DTEND;VALUE=DATE-TIME:19991201T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033104@cern.ch DESCRIPTION:The Globus project (http://www.globus.org) is developing basic software infrastructure for computations that integrate geographically di stributed computational and information resources. \n\nGlobus is a joint p roject of Argonne National Laboratory and the University of Southern Calif ornia's Information Sciences Institute. Led by Ian Foster and Carl Kesselm an\, it is the work of a talented project team at Argonne\, USC/ISI\, and the Aerospace Corporation\, with significant contributions also being made by other partners.\n\nThe talk will show how Globus can be used for the m anagement of on-line scientific data archives and for their on-demand remo te processing. Other Globus key points\, as the secure management of distr ibuted heterogeneous resources and the mapping of high level user requests into low-level systems procedures\, will be described. In \nthe talk we w ill also present how the Globus functionality is used in the SARA project (a joint project between CACR/Caltech and the HPC Lab/University of Lecce\ , http:// www.cacr. caltech.edu/sara)\, whose goal is to develop a Dynamic Earth Observation System (DEOS). \n\nSpeakers' Profile:\n\nGiovanni Alois io \nGiovanni Aloisio is Associate Professor of Information ProcessingSyst ems and Director of the High Performance Computing Laboratory of theDepart ment of Innovation Engineering (University of Lecce/ISUFI\, Italy). His re search interests include parallel and distributed computer architectures. Since 1986 to 1990 he contributed to the Caltech Concurrent Computation Pr ogram (C3P) led by Geoffrey C. Fox at the California Institute of Technolo gy\, investigating the efficiency of the Hypercube architecture in Real-Ti me SAR data processing. Actually he collaborates with the Center for Advan ced Computing Research (CACR) of the California Institute of Technology \n on High Performance Distributed Computing projects. He also collaborates w ith Carl Kesselman of the Information Sciences Institute\, on the use of ' Computational Grids' for the management of 'large collections of scientifi c data'.\n\nPaolo Falabella \nPaolo Falabella is a member of the High Perf ormance Computing Lab of the University of Lecce/ISUFI\, Italy. In July 19 99 he graduated 'cum laude' in Computer Engineering with Prof. G. Aloisio. His thesis was a study of the use of the GLOBUS toolkit to build a web-ba sed environment for the processing on demand of EOS images. He has been tr aining on Globus with Prof. Carl Kesselman at the Information Sciences Ins titute of Marina del Rey (Los Angeles\, California).Organiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033104 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033104 END:VEVENT BEGIN:VEVENT SUMMARY:Knowledge Management - Hype or Reality??\n DTSTART;VALUE=DATE-TIME:19991209T100000Z DTEND;VALUE=DATE-TIME:19991209T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033105@cern.ch DESCRIPTION:The explosion of data and information is dramatically changing the way an organisation makes decisions. It is no longer who you know\, b ut what you know. The more successful organisations will use knowledge as a precious commodity. Knowledge is becoming the new currency.\n\nAll compa nies\, from small independent firms to large corporate enterprises\, need to be able to manage and use knowledge to successfully compete in the fast paced global economy. Technological innovations will broaden capabilities and enable business to exploit the information they already have.\n\nCath y Lasser\, CIO\, Director of Information Technology for IBM Research will discuss Knowledge Management. What it is\, what it isn't and how it can im pact your organisation.\n\nAbout the speaker\n\nCathy joined IBM in 1978 a s a programmer supporting Test Engineering in Endicott\, New York. In 1982 \, Cathy joined the IBM Credit Corporation where she developed and managed the information centre and advanced technology development departments. I n 1993\, Cathy joined the PC Company where her organisation was responsibl e for executive \ninformation systems\, decision support systems and manuf acturing process support. She then came to IBM Research in 1996 as executi ve assistant to the Senior Vice President of Research. Cathy is also a mem ber of the Research Management Committee. Cathy holds a BS in Mathematics/ Computer Science from SUNY Binghamton and an MBA in Finance from Iona coll ege. Note: Please note the usual date and time \n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=a033105 LOCATION:CERN TH Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033105 END:VEVENT BEGIN:VEVENT SUMMARY:Web-Based Lectures - the Next Step DTSTART;VALUE=DATE-TIME:20000112T150000Z DTEND;VALUE=DATE-TIME:20000112T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033106@cern.ch DESCRIPTION:We have developed several lecture capture tools which easily\n allow educational material to be captured and delivered over the web.\nThe se tools include the Sync-O-Matic-3000 used to produce the CERN\nweb-lectu re archive and our newest tool\, the ClipBoard-2000 which is\ndesigned to be used with very little training or infrastructure. As the capture tools evolve in their usability\, new challenges arise. The media produced by today's capture tools needs to be reusable over a wide range of delivery m echanisms and connection capabilities. In the short-run\, we just generat e a number of differrent versions of each lecture\, or place several copie s of each lecture on several servers. Ultimately this must be automated w ith the data being served from a network of repositories\, allowing the us er to view the material using whatever tools they have at hand. This mean s a multi-technology capability for a lecture. Ultimately users should be able to get media on any operating system (PC\, Mac\, JUNIX) and using an y media technology (Real\, Microsoft\, Quicktime\, MPEG) and at any perfor mance level from a 386PC up to a high-end UNIX system. It is not possible to pre-build the hundreds or even thousands of different versions\nrequir ed to satisfy the variety of user requirements for viewing. This leads us to the need for dynamically generated content from the server at the mome nt that each user demands the information. CERN and the University of Mic higan have already collaborated on demonstrating the value of capturing an d storing web-based content using the Sync-O-Matic-3000. CERN and the Univ ersity of Michigan are starting to work on this repository problem togethe r. The presentation will discuss both the current state of the art in lec ture capture and the planned efforts in the lecture repository area. One interesting thread is to start work with the W3C to develop a standard for lecture interchange between repositories.\n\nReferences:\n\nAuthor page: http://www-personal.umich.edu/~csev/\nLecture Archive: http://webcast.cer n.ch/Projects/WebLectureArchive/\nOrganiser(s): G.Folger / ITNote: *** Ple ase note unusual PLACE ***\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=a033106 LOCATION:CERN 40-S2-A01 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033106 END:VEVENT BEGIN:VEVENT SUMMARY:Grid Computing: Resource Sharing and Coordinated Problem Solving i n Scalable Distributed Communities DTSTART;VALUE=DATE-TIME:20000117T150000Z DTEND;VALUE=DATE-TIME:20000117T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033109@cern.ch DESCRIPTION:Grid computing technologies enable controlled resource sharing in distributed communities and the coordinated use of those shared resour ces as community members tackle common goals. These technologies include n ew protocols\, services\, and APIs for secure resource access\, resource m anagement\, fault detection\, communication\, and so forth\, that in term enable new application concepts such as virtual data\, smart instruments\, collaborative design spaces\, and metacomputations. In this talk\, we rev iew applications that are motivating widespread interest in Grid concepts within the scientific and engineering communities. Then\, we describe the Globus Grid architecture that has been adopted by many Grid projects\, foc using in particular on our security\, resource management\, and data manag ement technologies.\n\nABOUT THE SPEAKERS\n\nDr. Ian Foster is Senior Scie ntist and Associate Director of the Mathematics and Computer Science Divis ion at Argonne National Laboratory\, Professor of Computer Science at the University of Chicago\, and Senior Fellow in the Argonne/U.Chicago Computa tion Institute. He has published four books and numerous articles in paral lel and distributed processing\, software engineering\, and computational science. He currently co-leads the Globus project with Dr. Carl Kesselman of USC/ISI\, which was awarded the 1997 Global Information Infrastructure 'Next Generation' award and which provides protocols and services used by many distributed computing projects worldwide. He co-founded the influenti al Grid Forum and recently co-edited with Kesselman a book on this topic\, published by Morgan-Kaufmann\, entitled 'The Grid: Blueprint for a New Co mputing Infrastructure.'Organiser(s): L. Pregernig\, IT/CE\n\nhttp://indic o.cern.ch/conferenceDisplay.py?confId=a033109 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033109 END:VEVENT BEGIN:VEVENT SUMMARY:Computational and data grids - are they important? DTSTART;VALUE=DATE-TIME:20000118T100000Z DTEND;VALUE=DATE-TIME:20000118T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033107@cern.ch DESCRIPTION:During 1999 it has become fashionable to talk about computatio nal and data grids. Sceptical audiences might wonder whether this renaming of 'distributed systems' is merely a passing fashion or represents someth ing more fundamentals. The speaker will argue that a fundamental change ha s indeed started. It is driven by the recognition that reliable high bandw idth wide-area networking has become generally available - an 'always-on' service - and that the quality and speeds available will continue to grow fast over the next few years. This permits communities to think of sharing computational tasks and access to data (at many different scales) in fund amentally new ways. While \n'the grids' will not replace all current modes of computing\, they will enable new applications as well as enhance many current ones.\n\nThe speaker will review several different grid initiative s and discuss the importance both of the quality of the network infrastruc ture and the way in which the scientific databases are structured.\n\nAbou t the speaker:\n\nDr. Paul Messina currently heads the Department of Energ y's Accelerated Strategic Computing Initiative (ASCI)\, which is concerned with developing the simulation capabilities needed to analyse \nand predi ct the performance\, safety\, and reliability of nuclear weapons and certi fy their functionality. \nAt the DoE he is on leave from Caltech\, where h e is Assistant VP for Scientific Computing and Director of the Centre for Advanced Computing Research. His research interests include advanced compu ter architectures\, especially their application to large-scale computatio ns in science and engineering and high-speed networks and computer perform ance evaluation.Note: * Please notes the unusual room\n\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=a033107 LOCATION:CERN 40-SS-D01 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033107 END:VEVENT BEGIN:VEVENT SUMMARY:MLPfit: a simple and powerful tool for training and using Multi-La yer Perceptrons\n DTSTART;VALUE=DATE-TIME:20000119T150000Z DTEND;VALUE=DATE-TIME:20000119T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033108@cern.ch DESCRIPTION:Artificial Neural Networks in general\, and Multi-Layer Percep trons in particular\, are now commonly used in many commercial\, industria l or research fields\, including High-Energy physics. \n\nIn this seminar\ , we will first describe some of the mathematical grounds of Multi-Layer P erceptrons (function approximation\, minimisation methods\, linear algebra \, etc.). We will then present the MLPfit package\, which:\n- can be used for both function approximation and classification tasks\n- implements pow erful minimisation methods in a fast and precise way - is available throug h various user interfaces\, including the latest version of PAW. Organiser (s): G. Folger / IT\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a 033108 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033108 END:VEVENT BEGIN:VEVENT SUMMARY:The application of semi-formal and formal techniques within the so ftware life cycle of safety \ncritical applications DTSTART;VALUE=DATE-TIME:20000202T150000Z DTEND;VALUE=DATE-TIME:20000202T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033110@cern.ch DESCRIPTION:The goal of this talk is to provide a short survey of various techniques\, which can be used within the software life cycle in the case of safety critical applications. Safety critical applications are applicat ions whose cost or possible impact on human lives requires the \ndevelopme nt of software of very high quality and as free from bugs as possible. The talk will describe some of the experiences done in the past 4-5 years by the group of Trento. \nWe will start by describing a small number of succe ss stories of previous industrial projects. For each project we will brief ly highlight the application area\, the problem solved\, the methodology a nd tools used.\n\nThe second part will provide more details about the meth odology and technology that we use:\nSoftware development and deployment. Some hints of UML might also be provided.\nValidation and verification. Af ter a brief introduction\, we will concentrate on the model checking techn ology. \nTime permitting\, we will provide a short description of NUMV\, a model checker (developed jointly with CMU) which now counts more than 50 installations world-wide.\nThe third line of research will highlight the l ines of research and technology development that we are currently pushing. \n\nAbout the speaker\nFausto Giunchiglia is professor of Computer Science (Artificial Intelligence and Software Engineering) at the University of T rento. He also collaborates with ITC-IRST (automated reasoning systems div isione)\, where he was Division head till November 1999.\nHe has a Degree in Electronic Engineering (1981) and a Doctorate (1987) (specialization in Computer Science) from the University of Genoa. He worked at Stanford Uni versity (Computer Science Dept.\, 1984-1987). He was a Research Fellow at the University of Edinburgh (Dept. of Artificial Intelligence\, 1987-1988) \, and a (part time) Visiting Fellow at Stanford University (Computer Scie nce Dept.\, 1990-1997).\n\nhttp://indico.cern.ch/conferenceDisplay.py?conf Id=a033110 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033110 END:VEVENT BEGIN:VEVENT SUMMARY:Highlights from CHEP 2000\n DTSTART;VALUE=DATE-TIME:20000216T150000Z DTEND;VALUE=DATE-TIME:20000216T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033112@cern.ch DESCRIPTION:The speakers will review the presentations given at this year' s Computing in High Energy Physics (CHEP 2000) conference in Padova (Italy )\, both from a general point of view and in their specific areas of inter est. Organiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=a033112 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033112 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
ORACLE DEVELOPERS FORUM\n DTSTART;VALUE=DATE-TIME:20000217T090000Z DTEND;VALUE=DATE-TIME:20000217T131500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033111@cern.ch DESCRIPTION:10.00 Oracle Web Server as a solution to integrate distinct Or acle-based information systems\, Pedro Martel / EST-ISS Division\n\nVery l arge projects require an information system built up of several distinct ( and most commonly not co-operating) subsystems\, each one specialised in i ts own scope. Contractual compliance verification becomes a very difficult task where the user must not only master \neach tool per se\, but also un derstand the context of data and information in all of them.\nThe 'Travell ers application' was developed as an Oracle Web Server based Manufacturing Follow Up Tool\, integrating several isolated Oracle based Engineering (C ADIM)\, Document Administration and Equipment Data (MP5)\n\nManagement sys tems.\nA generalised concept of a single user interface to federated datab ases has been tested in the Travellers application with very good results. \nAnother goal of the implementation was the enforcement of generalised bu siness rules rather at an inter-system level than that of a per system. Th e paper describes the underlying architecture and the first successful app lication of this concept at CERN.\n\n11.30 Update from Oracle OpenWorld\, Los Angeles\, Eric Grancher / IT-DB DivisionOrganiser(s): Pedro Martel / E ST Division \n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033111 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033111 END:VEVENT BEGIN:VEVENT SUMMARY:Searching multiple Web servers in HEP\n DTSTART;VALUE=DATE-TIME:20000223T150000Z DTEND;VALUE=DATE-TIME:20000223T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033116@cern.ch DESCRIPTION:Distributed search provides parallel searching across the hete rogeneous HEP-databases (eg. CERN Document Server\, SLAC and KEK). The too l called HEPDOC developed for this purpose will be discussed: how it was m ade and how to use it. There will also be an overview of other similar too ls.\nIt will be explained how this parallel searching technique could be a pplied in many areas where documents are distributed on different database s and servers.Organiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/confere nceDisplay.py?confId=a033116 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033116 END:VEVENT BEGIN:VEVENT SUMMARY:Testing Object-Oriented Software\n DTSTART;VALUE=DATE-TIME:20000308T150000Z DTEND;VALUE=DATE-TIME:20000308T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033113@cern.ch DESCRIPTION:Testing software is fraught with fundamental problems and diff iculties and object-oriented software\, contrary to initial popular belief \, has made the process even more difficult. This talk will briefly examin e some of the basic problems in testing and go on to illustrate the additi onal problems that object-oriented programming has introduced.\n\nA number of testing approaches are examined\, based on a variety of criteria (such as data flow and state)\, and the limitations of these approaches discuss ed. The talk concludes by considering some of the open problems and ways f orward.Organiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDisp lay.py?confId=a033113 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033113 END:VEVENT BEGIN:VEVENT SUMMARY:Archiving of electronics designs with EDMS\n DTSTART;VALUE=DATE-TIME:20000510T140000Z DTEND;VALUE=DATE-TIME:20000510T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033117@cern.ch DESCRIPTION:Equipment for the Large Hadron Collider (LHC) accelerator and experiments comes from many sources and will have a life cycle of 15-20 ye ars. It is therefore essential to keep track of designs and documentation throughout the lifetime of the LHC. To address these needs\, CERN has intr oduced an Engineering Data Management System (EDMS).Work is now under way to use EDMS also for electronics designs. The seminar will describe the pr oblems related to archiving of \nelectronics design data. This will be fol lowed by a brief introduction to the CERN EDMS CADIM/EDB. Furthermore the pilot project with an interface between Electronic Design Automation (EDA) tools and EDMS is presented. Finally there will be demo of how to access the data stored in EDMS via the Web and an outline of the next steps.\n\nA fter the seminar there will be a short demo of the EDA-EDMS interface for those who are interested.Organiser(s): G.Folger / IT\n\nhttp://indico.cern .ch/conferenceDisplay.py?confId=a033117 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033117 END:VEVENT BEGIN:VEVENT SUMMARY:The Trillian Project\n DTSTART;VALUE=DATE-TIME:20000517T140000Z DTEND;VALUE=DATE-TIME:20000517T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033118@cern.ch DESCRIPTION:CERN (via a small team in IT) has been working with several la rge computer companies (such as HP\, INTEL\, and Redhat) to adapt Linux to new Intel 64-bit architecture (IA-64)\, which is projected to replace alm ost completely the current 32-bit architecture by LHC start-up.\n\nThe pre sentation will review the first implementation of the IA-64 architecture\, the 'Itanium' processor\, and the Linux porting status. This covers both the kernel and the tool-chain. Since good compiler technology is paramount to exploiting the VLIW-like nature of IA-64\, the speaker will also addre ss the issue of porting CERN applications\, such as GEANT4 and SIXTRACK.\n \nOrganiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDisplay.p y?confId=a033118 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033118 END:VEVENT BEGIN:VEVENT SUMMARY:Quantum Computing (Defect-tolerant Molecular Nanoelectronics)\n DTSTART;VALUE=DATE-TIME:20000530T120000Z DTEND;VALUE=DATE-TIME:20000530T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033119@cern.ch DESCRIPTION:Both economic and general physics considerations indicate that the scaling era of the CMOS integrated circuit will saturate around the y ear 2010. However\, fundamental physical laws also indicate that it should be possible to compute with a power efficiency that is at least a billion times better than present silicon electronics. The most \nstraightforward way currently known to achieve such efficiencies are to make things very much smaller than they are at present.\n\nTwo complementary avenues of res earch are relevant to such future nanoelectronic systems: the development of a quantum-state switching device and the design of a system that will a ssemble itself once the discrete components have been brought together.\n\ nA group of researchers from UCLA and Hewlett-Packard Labs have recently d emonstrated that it is possible to construct molecular switches in a solid state device that can be set and read electronically\, although the switc h itself is nonreversible. Work is \ncontinuing in this area in order to m ake reversible switches\, as well as in the self-assembly of the nanowire arrays.\n\nVery real progress has been made on many different fronts\, but there are still significant opportunities and requirements for invention and discovery before nanoelectronics are a reality.\n\nAbout the speaker:\ nR. Stanley Williams is Senior Principal Laboratory Scientist and Director of the Quantum Structures Research Initiative (QSRI)\, the basic research department at Hewlett-Packard Laboratories.\nDr. Williams attended Rice U niversity from 1970-74\, where he obtained his B. A. degree in Chemical Ph ysics. He attended the University of California Berkeley from 1974-1978\, where he obtained his M. S. and PhD. degrees in Physical Chemistry\n\nhtt p://indico.cern.ch/conferenceDisplay.py?confId=a033119 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033119 END:VEVENT BEGIN:VEVENT SUMMARY:The MAP Project\n DTSTART;VALUE=DATE-TIME:20000531T140000Z DTEND;VALUE=DATE-TIME:20000531T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033120@cern.ch DESCRIPTION:The Monte Carlo Array Processor (MAP) at Liverpool University has been fully operational for 5 months. The system consists of 300 PC's c onnected via fast ethernet. In order to operate MAP a flow control system (FCS) was written. Details of the performance of MAP and of the FCS will b e discussed together with issues arising from the first 6 months of operat ion. Plans for the scaling of MAP to an LHC era system will be presented.O rganiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=a033120 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033120 END:VEVENT BEGIN:VEVENT SUMMARY:Experiences during the 2000 ALICE Data Challenge DTSTART;VALUE=DATE-TIME:20000607T140000Z DTEND;VALUE=DATE-TIME:20000607T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033121@cern.ch DESCRIPTION:The ALICE LHC experiment plans to acquire data at 1.25 GB/sec from 2005\, the highest rate of the four LHC experiments. Periodically ALI CE and IT collaborate to run a long simulation of their data acquisition a nd high level triggers from front-end data collectors through to tertiary storage. These 'Data Challenges'\, which involve the prototype of the ALIC E data acquisition system DATE\, the ALICE Off-line framework AliRoot\, th e new CERN data management software CASTOR and the commercial (IBM and Com paq) High Performance Storage System product (HPSS) intend to make step-wi se progress towards the final data acquisition and filtering environment. The challenge in 1999 reached 15MB/sec and the target rate for 2000 was 10 0MB/sec. This was reached for short periods and long periods of stable run ning at lower rates were demonstrated. The talk will describe the componen ts of the chain\, the experience gained at each level and the lessons lear ned that are both to the benefit of many CERN users and influencing the ne xt data challenge.Organiser(s): G.Folger / IT\n\nhttp://indico.cern.ch/con ferenceDisplay.py?confId=a033121 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033121 END:VEVENT BEGIN:VEVENT SUMMARY:A Vector Math Library for IA-64\n DTSTART;VALUE=DATE-TIME:20000614T140000Z DTEND;VALUE=DATE-TIME:20000614T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033115@cern.ch DESCRIPTION:Hewlett-Packard is preparing a Vector Math Library for the IA- 64 architecture to appear later this year. During 1998-99 the library was prototyped and validated in collaboration with CERN. The library deploys s everal new algorithms and exploits the 82-bit floating-point registers as well as the software pipelining capability of the new Intel architecture. Together these features offer exceptionally good accuracy\, very high spee d and full versatility across the whole range of arguments. The speaker wi ll discuss all these features in turn in order to demonstrate the value of this new mathematical library.\n\nhttp://indico.cern.ch/conferenceDisplay .py?confId=a033115 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033115 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
Web publishing using PL/SQL and Java DTSTART;VALUE=DATE-TIME:20000615T080000Z DTEND;VALUE=DATE-TIME:20000615T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033114@cern.ch DESCRIPTION:At CERN\, as increasingly elsewhere\, server based application s using the PL/SQL cartridge and Java are preferred as they make minimum d emands on the client and ease maintenance. The PL/SQL cartridge\, the orig inal method offered by Oracle\, has proved reliable and effective. We have used it to build web application for more than four years. Newer applicat ions are using servlets together with the application server.\n\nThis pape r presents the different techniques that we designed\, and the pitfalls th at we encountered in diverse projects such as a technology transfer databa se\, a product management tool and a system to keep track of physics event s.\n\nImportant considerations are security\, ease of maintenance\, transa ction handling and modularity. Most of the experience has been gained usin g the PL/SQL cartridge and this will be contrasted and compared with the n ewer Java techniques.Organiser(s): Pedro Martel / EST Division\n\nhttp://i ndico.cern.ch/conferenceDisplay.py?confId=a033114 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033114 END:VEVENT BEGIN:VEVENT SUMMARY:Part 1: Cjj: a subset of C++ compliant with Java\nPart 2: Reverse engineering of the UML Class Diagram from C++ Code in Presence of Weakly \ nTyped DTSTART;VALUE=DATE-TIME:20000712T140000Z DTEND;VALUE=DATE-TIME:20000712T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033123@cern.ch DESCRIPTION:Part 1\nThe C++ programming language offers a wide range of co ding alternatives to programmers. Some language characteristics inherited from C are potential sources of problems during the evolution of the syste m. Global variables\, arbitrary branches\, pointers\, memory management in structions and macro directives are examples of them. The resulting code m ay be difficult to understand\, maintain and test\, if such language featu res are used without proper discipline. The Java language was designed wit h the explicit goal of excluding all troublesome features that can be foun d in C++\, by disregarding some of them (e.g.\, global)\, and disciplining other ones (e.g.\, pointers). For several companies the migration to Java is not strategically feasible. Nevertheless\, the adoption of its philoso phy within C++ code could help improve the quality of the software. In thi s talk a subset of C++\, named Cjj\, is presented which is compliant with Java\, in that it \nenforces a Java-like programming style\, while forbidd ing all C++constructs that are in contrast with it. A tool for the verific ation of the compliance of an input source with Cjj was developed. \nIt ve rifies if any of the constraints defining Cjj is violated.\n\nPart 2\nThe Unified Modelling Language (UML) is becoming a standard for the design and development of object oriented systems. Its usage during the inception st ep of the development process allows describing the system at a high level \, thereby hiding irrelevant details and providing a common language for d iscussing about design rationale and alternative decisions. UML diagrams\, and in particular the most frequently used one\, the class diagram\, repr esent a valuable source of information even after the delivery of the syst em\, when it enters the \nmaintenance phase. Several tools provide a rever se engineering engine to recover it directly from the code. \nIn this talk \, an algorithm is proposed for the improvement of the accuracy of the UML class diagram extracted from the code. Specifically\, important informati on about inter-class relations may be missed in a reverse engineered class diagram\, when weakly typed containers\, i.e.\, containers collecting obj ects whose type is the top of the inheritance hierarchy\, are employed. In fact\, the class of the contained objects is not directly known\, and the refore no relation with it is apparent from the container declaration.\nTh e container type inference algorithm we propose is based on a flow insensi tive propagation of type information along all variables that can store a reference to the contained objects. Details are provided for its implement ation with reference to the C++ language. A tool was then developed for th e reverse engineering of the UML class diagram from C++ code\, which explo its the proposed algorithm.Organiser(s): G.Folger / ITNote: This seminar c onsists of two presentations of about 25 minutes each\n\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=a033123 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033123 END:VEVENT BEGIN:VEVENT SUMMARY:Linux Clustering and Storage Management\n DTSTART;VALUE=DATE-TIME:20000721T120000Z DTEND;VALUE=DATE-TIME:20000721T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033122@cern.ch DESCRIPTION:One of the most important Linux efforts in the year 2000 will be the \nbeginning of development of advanced storage management software. In \nthis talk we will review progress on file system and storage softwar e \nprojects.\n\nJournal file systems and logical volume management will h elp Linux to run on larger servers. Several projects are now close to comp letion and will provide much improved file system architecture.\n\nInterMe zzo is a replicating file system\, which is possibly suited to \nsynchroni se system files on systems in clusters\, or to synchronise storage on mobi le computers with servers.\n\nFinally we will also report on a novel objec t storage architecture\, called Lustre. Lustre allows storage management t o proceed through a stackable interface. Several file systems for Lustre a re under development.\n\nAbout the speaker:\nPeter J. Braam specialises in storage\, clustering and file system software and is best known for his w ork on the Lustre(TM) project and on the Coda and InterMezzo(TM) file syst ems. He received his PhD in 1987 and subsequently held tenured faculty pos itions at Utah and Oxford where he began to teach Computer Science in 1992 . He joined Carnegie Mellon's faculty in 1996. In 1999 he began to commit most of his time to the Linux industry as a part time Cluster and File Sys tems Architect for Red Hat. In March 2000 he joined TurboLinux as Chief Ar chitect.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033122 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033122 END:VEVENT BEGIN:VEVENT SUMMARY:High Performance Networking Workshop DTSTART;VALUE=DATE-TIME:20000915T070000Z DTEND;VALUE=DATE-TIME:20000915T080000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033382@cern.ch DESCRIPTION:9.00 Opening \nHans Hoffmann\, Director of Computing and Techn ology Transfer / CERN\n\n9.10 Grid Computing\nBen Segal / CERN\n\n9.40 CER N Activities: GSN-Optical\, GSN to OC48c\, Plans for 10 Gigabit Ethernet\, ST for LINUX.\nArie Van Praag / CERN\n\n10.00 ST Efficiency \nAlberto Gug lielmi / Genroco\, USA\n\n10.30 SAN Technologies: SCSI over ST \nGenroco\, USA\n\n11.00 'iSCSI - A SCSI mapping over TCP'\nJulian Satran / IBM\, Isr ael \n\n11.30 Cluster operating systems\, GFS\, SANergy\nDimitry Chernysho v / Tivoli\, USA\n\n12.00 CXFS\nSGI\, USA\n\n12.30 SAN Experts Facility Eu rope: the high speed \nJarno Puff / SAN Training Center\, networking refer ences point in Europe\, Italy\n\n12.40 LUNCH \n\n14.00 10 Gbit Ethernet \n Walther Dey / Cisco\n\n14.30 10 Gbit Ethernet chip technology \nFred Moham adi / Broadcom\, USA\n\n15.00 Infiniband \nBob Pearson / Vieo\, USA\n\n15. 30 to 18.00 Technology Demonstrations in the CERN Computer Centre\n\nFor t he technology demonstration it is planned to show:\n\n- The first GSN Opti cal connections\n- GSN switches from ESSENTIAL\, GENROCO and VIEO\n- PCI t o GSN NICs from ESSENTIAL\, GENROCO and VIEO\n- The new SGI Origin 3000 wi th native GSN NICs\n- Bridges equipped to connect GSN to Gigabit Ethernet\ , Fibre Channel and OC48c\n- Cluster File systems from SANergy and SGI\n- ST and SCSI-ST connections over Gigabit Ethernet.\n- ISCSI - SCSI over IP\ n\nHigh performance networking practice today offers Gigabit/sec media \ns peeds as well as protocol stacks capable of handling data rates of this or der while running on hardware that is affordable by the general user. At t he same time we are seeing the emergence of 10-Gigabit/sec technology\, pl us an explosion of interest in SAN's\, where the TCP/IP stack is replaced by more direct means of running storage protocols such as SCSI over Fibre Channel as well as other media.\n\nHighlights of the CERN Workshop are rep orts and/or demonstrations on:\n\nGSN: Gigabyte System Network\, the first ANSI standard 10-Gigabit \ntechnology\, running today on both parallel co pper and parallel optical fibre links. (the latter just developed at CERN) .\n\n10-GigE: the rapidly emerging standard for Ethernet at 10 Gigabits/se c.\n\nST Protocol: the Scheduled Transfer Protocol\, permitting both high \nbandwidth and low latency support including support for SCSI over \nstan dard media such as Gigabit Ethernet (GigE).\n\nSAN Bridges: SCSI-over-ST ( SST)\, running over various network media \nincluding GSN and GigE\, bridg ed to standard Fibre Channel storage.\n\nThe Workshop is organized in coll aboration with the High Performance \nNetwork Forum (HNF).Organiser(s): A. Van PraagNote: For more information and last minute updates contact A. Va n Praag at 75034\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033 382 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033382 END:VEVENT BEGIN:VEVENT SUMMARY:Global Computing Technology Outlook DTSTART;VALUE=DATE-TIME:20000928T090000Z DTEND;VALUE=DATE-TIME:20000928T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033124@cern.ch DESCRIPTION:Continued rapid evolutions of technologies to process\, store and communicate information are crossing thresholds that redefine how comp uting is exploited for both end-users and businesses. We will describe the basic trends for these base technologies\, with an eye on new ways they m ay be exploited as key thresholds are crossed.\nBeyond base technology\, w e will discuss the migration of high-performance computing hardware and ap plications from the scientific community into the core of the business com munity. This Deep Computing trend can not only move innovative business id eas into the marketplace very quickly\, it can also lead to a level of opt imisation in business performance that will redefine winners and losers in fields as large as the Life Sciences Industry.\n\nAbout the speaker:\nDav e joined IBM Research in 1988\, after receiving a Ph.D. in Low-Temperature Physics from Cornell University\, New York\, USA. He has held a variety o f technical and management positions within IBM Research\, in areas rangin g from high-performance interconnects to software technology and integrate d solutions. After serving as General Manager of Global Solutions in IBM's Government Industry sector\, he returned to IBM Research at the beginning of 1999 as Director of the Zurich Research Laboratory. Currently\, he is Vice-President\, Emerging \nBusiness.\n\nhttp://indico.cern.ch/conferenceD isplay.py?confId=a033124 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033124 END:VEVENT BEGIN:VEVENT SUMMARY:What comes next in Internet infrastructure DTSTART;VALUE=DATE-TIME:20001011T120000Z DTEND;VALUE=DATE-TIME:20001011T121500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033125@cern.ch DESCRIPTION:This presentation will consider how the network used to be how it is today\, and how it needs to be. Major issues are transparency\, add ressability\, and security and perceived service quality. The talk will di scuss whether there is an integration strategy for quality of service (QOS )\, what QOS solutions we have today\, and what is still missing for QOS. It will also cover the need for and prospects of IPv6. Finally it will con sider tactical solutions in use today.\n\nBio:\nBrian E. Carpenter is Prog ram Director\, Internet Standards and Technology\, for IBM. He is currentl y based at iCAIR\, the international Centre for Advanced Internet Research \, where he is also Chief Architect.\nPreviously he led the networking gro up here at CERN from 1985 to 1996. \nHe holds a first degree in physics an d a Ph.D. in computer science\, and is an M.I.E.E. He is an active partici pant in the Internet Engineering Task Force and a member of the Internet A rchitecture Board\, which he chaired from July 1995 to March 2000. He is c urrently Chairman of the Board of Trustees of the Internet Society\, and a member of the Board of the Unicode Consortium.\n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=a033125 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033125 END:VEVENT BEGIN:VEVENT SUMMARY:High Performance Networking for Data Grids DTSTART;VALUE=DATE-TIME:20001025T140000Z DTEND;VALUE=DATE-TIME:20001025T141500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033127@cern.ch DESCRIPTION:Large distributed collaborations like the EU (European Union) Data Grid \nProject will require the ability to move vast quantities of da ta across wide area networks (WAN) as efficiently and effectively as possi ble. However\, obtaining good TCP (Transmission Control Protocol) throughp ut across a wide area network usually requires some tuning. This is especi ally true in high-speed networks\, where\, even without congestion\, an ap plication might see only a small percentage of the total available bandwid th. \nThis talk will start with a brief introduction on how TCP works. Nex t\, it will describe analysis tools and techniques required to obtain good throughput. Finally\, the talk will cover some general methods for improv ing application performance in a high-speed WAN\, or 'Grid'\, environment. \n\nAbout the speaker: \nBrian L. Tierney is a Staff Scientist and group l eader of the Data Intensive Distributed Computing Group\, which is part of the Distributed Systems Department at Lawrence Berkeley National Laborato ry. His research interests include data intensive distributed computing\, high-speed I/O systems\, and distributed system performance monitoring and analysis. He is currently on leave at CERN for one year as a scientfic as sociate. Mr. Tierney has an M.S. in Computer Science from San Francisco St ate University\, and a B.A. in physics from the University of Iowa.Organis er(s): L. Pregernig\, IT/CE\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a033127 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033127 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
ORACLE DEVELOPERS FORUM DTSTART;VALUE=DATE-TIME:20001026T080000Z DTEND;VALUE=DATE-TIME:20001026T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033126@cern.ch DESCRIPTION:10:00 - Building a High Performance Java Servlet E-Business Ap plication \nUsing JServer and Oracle Workflow - Derek Mathieson - AS/IDS\n \nAt the dawn of the third millennium CERN is embarking on its largest \nc onstruction ever. Physicists world-wide require web-based access to genera te\, route and authorise a multitude of electronic documents\, such as pur chase orders\, transport requests\, with extremely tight deadlines. Eighte en months ago CERN embarked in the direction of Enterprise JavaBeans (EJB) architecture and a Java Servlet API. Now full production includes 1000 di fferent users per day on a multi-lingual system processing 300\,000 electr onic authorisations per year using Oracles high performance JServer Java V M in the heart of an Oracle8i database. This session walks through this su ccess story\, from the \nearly design decisions\, the promises of the EJB architecture\, client vs. server-side validation\, and the risks and pay-o ff of going production with such new technologies. The discussion includes practical advice from the chief developer of this project including the i mplications of each design decision\, and a critical comparison of alterna tive Java solutions. \n\n10:45 - Coffee break\n\n11:00 - Summary from Orac le OpenWorld 2000 - Jamie Shiers - IT/DBOrganiser(s): Pedro Martel / EST D ivision\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033126 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033126 END:VEVENT BEGIN:VEVENT SUMMARY:Presentations DTSTART;VALUE=DATE-TIME:20001101T150000Z DTEND;VALUE=DATE-TIME:20001101T151500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033129@cern.ch DESCRIPTION:Two fellows from IT's Computing Support for Engineering group will give the following presentations (appr. 30 minutes each):\n\n1) Elect romagnetic Simulation Tools: from Accelerators to Detectors.\nby Tomas Mot os Lopez\, CERN/IT\n\nThe design process of LHC involves new achievements in both the accelerator and detector technologies. Various commercial elec tronic design automation (EDA) programs are now available to address these problems. These include 3-D full wave electromagnetic field solvers used to characterise a great amount of different subsystems: particle \ndetecto rs\, high-speed connectors\, transmission line structures\, and resonant c avities. This presentation will show how these tools are used at CERN and how they can interrelate with other engineering tools like PSpice or Matla b.\n\nABOUT THE SPEAKER: \nTomas Motos Lopez holds a Telecommunications En gineering degree from the Polytechnic University of Valencia. He began his professional career at Alcatel Telecom Spain\, as a digital designer. Sin ce 1999 he works as a fellow of the IT division. His current interests inc lude electromagnetic simulation\, signal integrity\, and digital design.\n \n2) High-Level System Design using Foresight\nby Giovanna DiMarzo-Serugen do\, CERN/IT\n\nThe Foresight tool is a system-level modelling environment that allows one to capture\, design\, model\, and simulate (execute) syst em specifications. With the help of this tool\, designers can identify and eliminate early in the product development process errors\, such as\, con flicting requirements\, functional and performance problems\, etc.. The pr esentation will describe how to specify systems with Foresight and how to execute (simulate) system models. A practical example\, the \napplication of the tool to the ALICE DAQ (data-acquisition) system\, will conclude the presentation.\n\nABOUT THE SPEAKER: \nGiovanna Di Marzo Serugendo is curr ently working as a fellow of the IT division on the specification and simu lation of the ALICE DAQ System. Her interests are in software engineering and formal methods for distributed systems and multi-agent systems. Mrs Di Marzo Serugendo holds a M.Sc. in Mathematics\, a M.Sc. in Computer Scienc e from the University of Geneva\, and a Ph.D. in Software Engineering from the EPFL (Ecole Polytechnique Federale\, Lausanne)Organiser(s): L. Preger nig\, IT/CE\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033129 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033129 END:VEVENT BEGIN:VEVENT SUMMARY:Seminar Overview DTSTART;VALUE=DATE-TIME:20001108T150000Z DTEND;VALUE=DATE-TIME:20001108T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033128@cern.ch DESCRIPTION:As part of the evolution from mainframes to distributed comput ing CERN developed its own mass storage data management system\, SHIFT\, w hich has successfully managed LEP and fixed target data over many years. T he requirements of LHC go well beyond the original design of SHIFT and th ree years ago we adopted a two line approach namely to try a commercial so lution\, the High Performance Storage System (HPSS) from IBM\, and to deve lop a new Cern Advanced Storage (CASTOR) Manager that would first be used for the next generation of fixed target experiments then extended for LHC. When we started HPSS Compaq Alpha processors were our main tape servers so we included a joint project with a Compaq funded fellow to port the neede d parts of HPSS to Compaq while the authors of the SHIFT system took a CER N fellow to work on CASTOR. Both fellows report on their work in this semi nar.\n\n1) HPSS 4.1.1 on Compaq Tru64 Unix\nby Patrice Calegari\, CERN/IT\ n\nThe High-Performance Storage System (HPSS) provides hierarchical storag e management and services for the largest storage environments in the worl d. It was developed by laboratories and companies in the USA. At CERN\, it has been in use since 1998 to store user and experiment data. HPSS evolve s quickly and CERN participates in its development on Alpha processors. Th is presentation will show the improvements that were done this year by por ting the last HPSS versions to Compaq Tru64 Unix. An analysis of performan ce and quality achieved in heterogeneous environments will conclude the pr esentation.\n\nABOUT THE SPEAKER\nPatrice Calegari\, currently working as a fellow in CERN'S IT Division\, holds a Magistere (Master's Degree) in Co mputer Science from the 'Ecole Normale Supérieure de Lyon' and a Ph.D. in Computer Science (parallel computing) from the Swiss Federal Institute of Technology\, Lausanne (EPFL). His interests include parallel computing\, data management and software engineering.\n\n2) Experience with the CASTOR Mass Storage System: Practical\, Design and Development Issues\nby Jean-D amien Durand\, CERN/IT\n\nThis talk will focus on the philosophy behind su ch a project\, the effort and decisions involved in the design (scalabilit y\, modularity\, HSM functionality\, administration\, ...)\, development i ssues (programming language\, portability\, appplication-program interface \, bindings\, ...)\, and practical issues: how does it work.\n\nABOUT THE SPEAKER\nJean-Damien Durand\, currently working as a fellow in CERN'S IT D ivision\, holds a Ph.D. in Physics. He first got involved in computing wit h the DELPHI experiment\, as software librarian for two years. He then mov ed to IT-PDP section where he has been working since 1998 on SHIFT and CAS TOR.Organiser(s): L. Pregernig\, IT/CE\n\nhttp://indico.cern.ch/conference Display.py?confId=a033128 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033128 END:VEVENT BEGIN:VEVENT SUMMARY:Object Oriented Ntuple/Tag Analysis in Anaphe/LHC++ DTSTART;VALUE=DATE-TIME:20001115T150000Z DTEND;VALUE=DATE-TIME:20001115T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033130@cern.ch DESCRIPTION:The use of object oriented (OO) technologies and methodologies in High-energy Physics has been strongly increasing during the last few y ears. At the last HepVis'99 workshop\, a working group (AIDA) has been for med to systematically design a set of Abstract Types for various component s of data analysis tools. The use of these Abstract Types combined with dy namically loadable libraries allows the creation of extremely flexible and powerful data analysis tools.\nA set of interfaces and class libraries ha s been developed for analysis of Ntuple- (Tag-) like data structures. Thes e libraries are part of the Anaphe/LHC++ project and can be used either 's tandalone'\, e.g.\, in batch programs as well as in (AIDA-compliant) inter active analysis environments. We will present the design criteria for this project\, describe the underlying architecture\, and show its actual stat us. An outlook on the project's future development will conclude the prese ntation.\n\nABOUT THE SPEAKER\n\nZsolt Molnar\, currently working as a fel low in CERN's IT Division\, holds a degree in Theoretical Physics. Before joining CERN\, he has been working on GIS (Geographical Information System ) projects for the European Space Agency and on telecom software projects for Nokia and Symbian.Organiser(s): L. Pregernig\, IT/CE\n\nhttp://indico. cern.ch/conferenceDisplay.py?confId=a033130 LOCATION:CERN URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033130 END:VEVENT BEGIN:VEVENT SUMMARY:Next Generation HTML Rendering in GNOME DTSTART;VALUE=DATE-TIME:20001122T150000Z DTEND;VALUE=DATE-TIME:20001122T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033132@cern.ch DESCRIPTION:CodeFactory is developing Gtkhtml2\, the next generation HTML rendering widget for the GNU-Image-Manipulation-Package Toolkit (GTK)\, us ed in the GNOME desktop environment. It provides a simple\, lightweight me chanism for rendering and displaying structured text - specifically HTML a nd XHTML - in applications. By utilizing true Model-View-Controller (MVC) design and flexible event management using a DOM2 (Document Object Model) compliant engine it makes for an easily embeddable and powerful widget for use in modern applications in Linux\, and other UNIX-like operating syste ms.\n\nABOUT CodeFactory\n\nCodeFactory AB is an 'Open Source' consulting firm based in Sweden\, formed in early 2000. The company focuses on aiding companies and organizations with the development and use of open source s oftware\, while actively contributing development efforts back to the open source community itself.Organiser(s): L. Pregernig\, IT/CE\n\nhttp://indi co.cern.ch/conferenceDisplay.py?confId=a033132 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033132 END:VEVENT BEGIN:VEVENT SUMMARY:Integration of LDAP into the Messenging Infrastructure at CERN DTSTART;VALUE=DATE-TIME:20001129T150000Z DTEND;VALUE=DATE-TIME:20001129T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033131@cern.ch DESCRIPTION:LDAP\, the Lightweight Directory Access Protocol\, is a softwa re protocol for enabling anyone to locate individuals and other resources\ , such as files and devices in a network. The presentation will start with a description of the basic principles of LDAP\, such as its role as both an access protocol and a data structure. Next\, the presentation will high light the differences between a classical database and a directory service such as LDAP and describe various related concepts\, like object classes and schema definitions. After this introduction\, the integration of LDAP into CERN's messenging infrastructure will be \nshown. This ranges from th e simple global addressbooks to more powerful and complex areas\, such as storing distribution list configurations and profiles\, as well as mail an d web-based authentication. Future projects\, like message routing under s endmail and storing system-level-authentication information via the PAM/LD AP features offered under Unix shall also be discussed.\n\nAbout the speak er:\n\nRay Jackson\, currently working as a fellow in CERN's IT Division\, holds a first-class bachelors degree (with honours) in computer science f rom 'Manchester Metropolitan University' in the UK. After working for a ye ar in the PS division as an undergraduate\, he returned to CERN in June 19 99 to work in the IT Division's Internet Services group\, \nprimarily look ing after the messenging infrastructure and directory services. His main i nterests are LDAP\, sendmail\, and the development of Web applications.\n\ n\nOrganiser(s): \nLudwig Pregernig\nComputing Seminars /\nIT Department\n \n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033131 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033131 END:VEVENT BEGIN:VEVENT SUMMARY:Grid Computing: Resource Sharing and Coordinated Problem Solving i n Scalable Distributed Communities DTSTART;VALUE=DATE-TIME:20010117T150000Z DTEND;VALUE=DATE-TIME:20010117T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033133@cern.ch DESCRIPTION:Grid computing technologies enable controlled resource sharing in distributed communities and the coordinated use of those shared resour ces as community members tackle common goals. These technologies include n ew protocols\, services\, and APIs for secure resource access\, resource m anagement\, fault detection\, communication\, and so forth\, that in term enable new application concepts such as virtual data\, smart instruments\, collaborative design spaces\, and metacomputations.\n\nIn this talk\, we review applications that are motivating widespread interest in Grid concep ts within the scientific and engineering communities. Then\, we describe t he Globus Grid architecture that has been adopted by many Grid projects\, focusing in particular on our security\, resource management\, and data ma nagement technologies.\n\nAbout the speakers\n\nDr. Ian Foster is Senior S cientist and Associate Director of the Mathematics and Computer Science Di vision at Argonne National Laboratory\, Professor of Computer Science at t he University of Chicago\, and Senior Fellow in the Argonne/U.Chicago Comp utation Institute. He has published four books and numerous articles in pa rallel and distributed processing\, software engineering\, and computation al science. He currently co-leads the Globus project with Dr. Carl Kesselm an of USC/ISI\, which was awarded the 1997 Global Information Infrastructu re 'Next Generation' award and which provides protocols and services used by many distributed computing projects worldwide. He co-founded the influe ntial Grid Forum and recently co-edited with Kesselman a book on this topi c\, published by Morgan-Kaufmann\, entitled 'The Grid: Blueprint for a New Computing Infrastructure.'\n\nOrganiser(s): L. Pregernig\, IT/CE\n© CERN 2005 -\nMiguel Angel Marquina\n/ IT Department\n\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=a033133 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033133 END:VEVENT BEGIN:VEVENT SUMMARY:Design Strategies for Irregularly Structured Parallel Applications DTSTART;VALUE=DATE-TIME:20010214T150000Z DTEND;VALUE=DATE-TIME:20010214T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032795@cern.ch DESCRIPTION:The success of parallel computing in solving real-life computa tionally intensive problems relies on their efficient mapping and executio n on large-scale multiprocessor architectures. Many important applications are both unstructured and dynamic in nature\, making their efficient para llel implementation a daunting task.\n\nIn this talk I will present the pa rallelization of two irregularly structured computations: the Conjugate Gr adient algorithm and unstructured mesh adaptation. Both codes were develop ed with an MPI message passing implementation on the Cray T3E and the SGI Origin2000\, a shared-memory implementation using the cache coherent nonun iform memory access(CC-NUMA) of the Origin2000\, and a multithreaded versi on on the newly released Tera Multithreaded Architecture (MTA). A comparis on is presented between the critical factors of these parallel code develo pments\, including runtime\, scalability\, programmability\, and memory ov erhead. The overall results demonstrate that multithreaded systems offer a tremendous potential for quickly and efficiently solving some of the most challenging real-life problems on parallel computers.\n\nThis work won th e Best Paper Award at Supercomputing99.\n\nAbout the speaker:\n\nLeonid Ol iker is currently a computer scientist in the Future Technologies Group at the National Energy Research Scientific Computing Center (NERSC)\, locate d at Lawrence Berkeley National Laboratory (LBNL). In 1991\, he received a B.S.E. in computer engineering from the University of Pennsylvania and a B.S. in finance from the Wharton School of Business. He received his Ph.D. in computer science from the University of Colorado in 1998. Dr. Oliker h as held positions as a visiting researcher and a postdoctoral scientist at the Research Institute for Advanced Computer Science (RIACS) at NASA Ames Research Center. His research interests include the study of dynamically adapting algorithms on advanced parallel architectures\, job scheduling fo r effective system performance\, generic programming for scientific comput ations\, processor-in-memory clusters\, and resource management for mobile computing.\n\nOrganiser(s): L. Pregernig\, IT/CE\n© 2005 -\nMiguel Angel Marquina\n/ IT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py ?confId=a032795 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032795 END:VEVENT BEGIN:VEVENT SUMMARY:ObjectGlobe: Open Distributed Query Processing Services on the Int ernet DTSTART;VALUE=DATE-TIME:20010321T150000Z DTEND;VALUE=DATE-TIME:20010321T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033136@cern.ch DESCRIPTION:\nWe present the design of ObjectGlobe\, a distributed and ope n query processor for Internet data sources. The goal of the ObjectGlobe p roject is to establish an open market place in which data and query proces sing capabilities can be distributed and used by any kind of Internet appl ication. Furthermore\, ObjectGlobe integrates cycle providers (i.e.\, mach ines) which carry out query processing operators. The overall picture is t o make it possible to execute a query with --in principle-- unrelated quer y operators\, cycle providers and data sources.\n\nOne of the main challen ges in the design of such an open system is to ensure privacy and security . We discuss the ObjectGlobe security requirements\, show how basic compon ents such as the optimizer and runtime system need to be extended. Another challenge is quality of service (QoS) management so that users can constr ain the costs and running times of their queries.\n\nAbout the speakers:\n \nAlfons Kemper has been a Full Professor at the University of Passau sinc e 1993. He received his M. Sc. and Ph. D. degrees in Computer Science from the University of Southern California in Los Angeles in 1981 and 1984. Hi s research concentrates on optimization of advanced database systems and a pplications (e.g. data warehouses and decision-support query processing)\, on distributed data base systems and query processing\, and on complex da tabase application system performance analysis and tuning\, in particular for the SAP R/3 system. Prof. Kemper has authored over 75 publications inc luding three books on database systems.\n\nReinhard Braumandl (University of Passau) studied Computer Science at the University of Passau from 1992 to 1997. Since 1997\, he has been a research assistant in the department f or mathematics and computer science at the University of Passau. He is cur rently pursuing the Ph.D. degree. His research interests include query pro cessing\, performance analysis and QoS management in object-oriented and d istributed databases.\n\nOrganiser(s): L. Pregernig\, IT/CE\n© 2005 -\nMi guel Angel Marquina\n/ IT Department\n\n\nhttp://indico.cern.ch/conference Display.py?confId=a033136 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033136 END:VEVENT BEGIN:VEVENT SUMMARY:High Performance Networking as Sign of its Time DTSTART;VALUE=DATE-TIME:20010509T140000Z DTEND;VALUE=DATE-TIME:20010509T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033135@cern.ch DESCRIPTION:After an historical introduction about what people saw as high performance message communication in their time\, a far more technical pa rt follows that compares the three 10 Gbit/s network technologies that exi st or emerge. Some remarks will be made concerning the understanding of ne twork storage.\n\nABOUT THE SPEAKER\n\nArie van Praag has been working on High Performance networking since 1989. He has made major contributions to the standardisation process of a 10 Gbit/s Network\, including the input from research at CERN\, and. He was specially involved in the standardisat ion of a fibre optics standard for GSN and conversion standards for Gigabi t ethernet\, Scheduled Transfer\, and SONET/SDH. He is a member of the 'bo ard of directors' of the 'High Performance Networking Forum' HNF and chair man of HNF-Europe.Organiser(s): L. Pregernig\, IT/CE\n\nhttp://indico.cern .ch/conferenceDisplay.py?confId=a033135 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033135 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
ORACLE DEVELOPERS FORUM DTSTART;VALUE=DATE-TIME:20010516T080000Z DTEND;VALUE=DATE-TIME:20010516T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033134@cern.ch DESCRIPTION:10:00 - CAST Software (development and application mining tool s)\nStephan Petit - EST/ISS\n\nCAST Software proposes a set of tools aimed at software development and application mining. \nI will present these to ols that are currently used by the EST/ISS group in an Oracle PL/SQL envir onment: \n\n- SQL Builder: software development \n- Release Builder: softw are deployment \n- Application Viewer: application mining \n\nTheir most i mportant features will be highlighted. Better quality and faster developme nt brought by these integrated tools will be discussed. \n\n10:45 - Coffee break\n11:00 - Oracle Servers Configuration Update - Nilo Segura - IT/DB\ n\nOrganiser(s): Pedro Martel / EST Division\n\nhttp://indico.cern.ch/conf erenceDisplay.py?confId=a033134 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033134 END:VEVENT BEGIN:VEVENT SUMMARY:EP COMPUTING FORUM (RESERVED TO CERN-EP)
Migration to Windows 2 000 DTSTART;VALUE=DATE-TIME:20010523T120000Z DTEND;VALUE=DATE-TIME:20010523T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033137@cern.ch DESCRIPTION:14:00 The W2000 Migration task-force\, F.Hemmer (IT)\n\n14:30 Computing for engineering\, F.Hemmer on behalf of C.Eck\n\n14:50 W2000 for physics\, J.Closier (EP\, LHCb) \n\n15:15 Windows for CMS cristal calibra tion\, A.Branson (EP\, CMS)\n\n15:35 W2000 for Secretariat\, V.Renoult (EP \, LHCb/ALEPH)\n\n15:50 Discussion\, all\n\nOrganiser(s): Ph.Charpentier / EP\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033137 LOCATION:CERN 40-S2-B01 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033137 END:VEVENT BEGIN:VEVENT SUMMARY:New Features in Oracle 9i (Part 1) DTSTART;VALUE=DATE-TIME:20010625T120000Z DTEND;VALUE=DATE-TIME:20010625T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033139@cern.ch DESCRIPTION:The launch of Oracle 9i\, the latest release of Oracle's Objec t-Relational Database management system\, will take place during the Oracl e OpenWorld conference in Berlin June 18 - 21. These two seminars\, based upon presentations from OpenWorld\, cover new features in Oracle 9i in its support for XML\, Java and C++. Copies of the OpenWorld abstracts are inc luded below. Oracle9i has rapidly evolved into a database for all your dat a\, from simple to complex types. Multimedia data types such as images\, m aps\, video clips and audio clips are now common in many Web-based applica tions. Other software solutions also need to store data dealing with compl ex financial instruments\, engineering diagrams or molecular structures. W ith Oracle9i\, Oracle is completing the database object-relational vision by supporting full object modeling capabilities\, including inheritance an d multilevel collections\, with type evolution capabilities. This session presents the latest object-relational advances\, focusing on the core serv er features at a simple conceptual level and on the various language inter faces useful for manipulating server objects. The session also discusses c omponent frameworks\, SQL:1999\, emerging W3C XML standards\, how they map to these advances and more. New in Oracle9i\, Oracle C++ Call Interface ( OCCI) is a better OCI (Oracle Call Interface) for C++ programmers. With a extensive collection of easy-to-use C++ classes and templates\, OCCI provi des high-performance and comprehensive C++ interface to powerful Oracle9i object-relational and relational database features. OCCI is based on the m ost recent C++ standard along with its Standard Template Library (STL). In addition\, the OTT (Object Type Translator) utility has also been enhance d to generate C++ interface classes from user-defined types in the databas e to greatly simplify the development task. Attendees will learn to use OC CI for developing high-performance database applications.\n\nOrganiser(s): J. Shiers\, IT/DB\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a0 33139 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033139 END:VEVENT BEGIN:VEVENT SUMMARY:New Features in Oracle 9i (Part2) DTSTART;VALUE=DATE-TIME:20010625T140000Z DTEND;VALUE=DATE-TIME:20010625T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033140@cern.ch DESCRIPTION:Organiser(s): J. Shiers\, IT/DB\n\nhttp://indico.cern.ch/confe renceDisplay.py?confId=a033140 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033140 END:VEVENT BEGIN:VEVENT SUMMARY:The GridPort Architecture for Building Grid Portals DTSTART;VALUE=DATE-TIME:20010627T140000Z DTEND;VALUE=DATE-TIME:20010627T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033138@cern.ch DESCRIPTION:The Grid' is the term applied to the infrastructure being deve loped to interconnect highly distributed resources (compute\, archival\, i nstrumentation\, etc.) into a large\, parallel\, computational resource. G rid portals are emerging as a highly convenient mechanism for providing th e general scientific community with a familiar and simplified web-based in terface to the Grid and Grid services via the Web. The rapid increase in p ortal projects has led to a proliferation of solutions to the portal probl em. While each solution is of interest\, there is a need for protocols\, i nteroperable standards\, and common architectures to promote the sharing o f information\, data\, and resources. Our experiences in implementing mult iple computational science Grid portals at NPACI has led to the definition of a unique\, layered architecture for Grid portals that sits on the top of the middle tier of the classic 3-tiered Grid architecture and also inte racts with the collective layer of the Grid. Our architecture is an extens ion of these concepts\, and one that we believe will help to promote inter operability among Grid portals and portal-to-portal communication. The fea sibility of this architecture has been successfully demonstrated with the implementation of several production portals. In this talk we will describ e the toolkits we have built based on this architecture\, and demonstrate the usefulness of this architecture with a discussion of examples of sever al production portals.\n\nOrganiser(s): L. Pregernig\, IT/CE\n\nhttp://ind ico.cern.ch/conferenceDisplay.py?confId=a033138 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033138 END:VEVENT BEGIN:VEVENT SUMMARY:Grid Technologies and Applications: Architecture and Achievements DTSTART;VALUE=DATE-TIME:20010827T140000Z DTEND;VALUE=DATE-TIME:20010827T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032796@cern.ch DESCRIPTION:The last 18 months have seen significant advances in Grid comp uting\, both within and outside high energy physics. While in early 2000\, Grid computing was a novel concept that many were being exposed to for th e first time\, we now see considerable consensus on Grid architecture\, a solid and widely adopted technology base\, major funding initiatives\, a w ide variety of projects developing applications and technologies\, and maj or deployment projects aimed at creating robust Grid infrastructures. I pr ovide a summary of major developments and trends\, focusing in particular on the Globus open source Grid software project and the GriPhyN data grid project.\n\nABOUT THE SPEAKER\n\nDr. Ian Foster is Senior Scientist and As sociate Director of the Mathematics and Computer Science Division at Argon ne National Laboratory\, Professor of Computer Science at the University o f Chicago\, and Senior Fellow in the Argonne/U.Chicago Computation Institu te. He has published four books and numerous articles in parallel and dist ributed processing\, software engineering\, and computational science. He currently co-leads the Globus project with Dr. Carl Kesselman of USC/ISI\, which was awarded the 1997 Global Information Infrastructure 'Next Genera tion' award and which provides protocols and services used by many distrib uted computing projects worldwide. He co-founded the influential Grid Foru m and recently co-edited with Kesselman a book on this topic\, published b y Morgan-Kaufmann\, entitled 'The Grid: Blueprint for a New Computing Infr astructure'.\nOrganiser(s): L. Pregernig\, IT/US\n\nhttp://indico.cern.ch/ conferenceDisplay.py?confId=a032796 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032796 END:VEVENT BEGIN:VEVENT SUMMARY:Rules of Thumb in Data Engineering DTSTART;VALUE=DATE-TIME:20010829T120000Z DTEND;VALUE=DATE-TIME:20010829T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033142@cern.ch DESCRIPTION:This talk reexamines the rules-of-thumb for the design of data storage systems. Briefly\, it looks at storage\, processing\, and network ing costs\, ratios\, and trends with a particular focus on performance and price/performance. Amdahl's ratio laws for system design need only slight revision after 35 years - the major change being the increased use of RAM . An analysis also indicates storage should be used to cache both database and web data to save disk bandwidth\, network bandwidth\, and people's ti me. Surprisingly\, the 5-minute rule for disk caching becomes a cache-ever ything rule for web caching.\n\nBiography:\n\nJim Gray is part of Microsof t's research group. His work focuses on databases and transaction processi ng. Jim is active in the research community\, is an ACM\, NAE\, NAS\, and AAAS Fellow\, and received the ACM Turing Award for his work on transactio n processing. He is also a member of the PITAC\, and edits of a book serie s of books on data management.\n\nhttp://indico.cern.ch/conferenceDisplay. py?confId=a033142 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033142 END:VEVENT BEGIN:VEVENT SUMMARY:Federated Information for Data Grids DTSTART;VALUE=DATE-TIME:20010919T140000Z DTEND;VALUE=DATE-TIME:20010919T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033143@cern.ch DESCRIPTION:The internet has raised awareness of the need for metadata whi ch describes data held in information resources attached to networks so th at the data can be used in federated environments. This is becoming more i mportant with the advent of the high speed networks of the GRID which will enable disparate data to be linked for analytic purposes. The talk will c oncentrate on the presenter's interest in metadata\, which started in the early days of distributed databases and has persisted into its role in cre ating federated knowledge systems supporting different user's integration views for data in web environments where the data to be linked and fused c an be held in structured and unstructured collections. This will lead to t he evolving role of metadata in the future GRID technologies covering its role in data quality ensurance\, data curation and secondary analysis.Abou t the speakerAlex GRAY is Professor of Advanced Information Systems at Car diff University. He leads the OKS{Object and Knowledge Based Systems} rese arch group in the Department of Computer Science. He is a member of many c onference programme committees such as VLDB\, ICDE\, EDBT\, BNCOD\, etc. a nd has published more than 150 papers on distributed database systems\, di stributed information systems\, knowledge based systems. He holds a number of current grants in the areas of interoperability of information systems from EC\, EPSRC\, BBSRC and other funding bodies. His recent work has bee n concerned with building federated knowledge based systems for biodiversi ty information systems and engineering information systems. Another major interest is the distributed support of design systems for concurrent engin eering. He is involved in the recently announced Cardiff Regional GRID ce ntre funded by the DTI and EPSRC. In this centre he will be looking at how metadata can be used to improve the handling of data used in experimentat ion in the escience world. Two recent grants are concerned with handling d ata for magnetic wave experimentation in a grid environment and the creati on of a grid biodiversity demonstrator.Organiser(s): L. Pregernig\, IT/CE\ n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033143 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033143 END:VEVENT BEGIN:VEVENT SUMMARY:Highlights from CHEP 2001\n DTSTART;VALUE=DATE-TIME:20011010T140000Z DTEND;VALUE=DATE-TIME:20011010T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033141@cern.ch DESCRIPTION:David Myers/IT-CO\, Vlado Bahyl/IT-PDP\, Jean-Philippe \nBaud/ IT-PDP\, Harry Renshall/IT-PDP\, Gabriele Cosmo/IT-API\n\nThe speakers wil l review the presentations given at this year's \nComputing in High Energy Physics (CHEP 2001) conference in Beijing\, both from a general point of view and in their specific areas of interest.Organiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033141 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033141 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
ORACLE DEVELOPERS FORUM DTSTART;VALUE=DATE-TIME:20011015T080000Z DTEND;VALUE=DATE-TIME:20011015T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033145@cern.ch DESCRIPTION:10:00 Building a financial data warehouse with Oracle - Michae l Leif Angberg - AS/IDS\nWithin the CERN organization there are almost 100 0 budget holders\, \nlocated at CERN as well as at any of the 500 institut es around the world that CERN collaborates with. In order to develop a fin ancial decision support system for this type of community one must of cour se take into account a vast range of requirements\, but they can be accura tely summarized with the following 3 words:\n\n- Versatility\n\n- Speed\n\ n- Availability\n\nAn application meeting these requirements is currently being developed at CERN\, using Data Warehousing concepts offered by Oracl e\, as well \nas Oracles J2EE platform. The building of the data warehouse \, optimized for analytical querying\, makes use of Materialized Views\, P artitioned tables and adapted Query Rewrite to achieve a high level of ver satility on the database side.\nThe business logic/middle tier is implemen ted using a framework \nsimilar to the component-like EJB specification.\n (This presentation is part of EOUG 2001 conference)\n\n10:45 - Coffee brea k\n\n11:00 - Oracle Servers Configuration Update - Eric Grancher - IT/DBOr ganiser(s): Pedro Martel / EST Division\n\nhttp://indico.cern.ch/conferenc eDisplay.py?confId=a033145 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033145 END:VEVENT BEGIN:VEVENT SUMMARY:CVS Services for ATLAS DTSTART;VALUE=DATE-TIME:20011024T140000Z DTEND;VALUE=DATE-TIME:20011024T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033144@cern.ch DESCRIPTION:In many large software projects\, CVS [Concurrent Versions Sys tem] is the tool of choice for version control. Local access to the CVS re pository does not scale\; a client/server implementation is required. I wi ll describe the CVS server that has been set up for the ATLAS experiment a t CERN. It has been built of standard hardware and software components\, a nd has been designed such that serving other projects from the same server is easily possible. Recently\, the CVS repository of the CHORUS experimen t has been moved to the same CVS server. I will cover both the services of fered to the users\, and the implementation from a provider's point of vie w. \n\nAbout the speaker\n\nHelge Meinhard has been a staff member in CERN 's EP Division since 1994\, and has since worked on software\, software su pport and computing infrastructure for the CHORUS and ATLAS experiments. O rganiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/conferenc eDisplay.py?confId=a033144 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033144 END:VEVENT BEGIN:VEVENT SUMMARY:ORACLE DEVELOPERS FORUM:
Adding Mathematica software to an ORAC LE database DTSTART;VALUE=DATE-TIME:20011105T090000Z DTEND;VALUE=DATE-TIME:20011105T103000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033146@cern.ch DESCRIPTION:Scientific database often needs graphical and/or computational \nsoftware for sophisticated data query and presentation. The Internation al Gravitational Events Collaboration Database (IGECDB) stores gravitation al events from five gravitational antennas in the \nworld using a 3 tiers architecture. In IGECDB both needs have been \nsatisfied by using Wolfram Research Mathematica software interfaced \nvia standard (JMLink) and custo m Java classes in the middle tier. We \npresent the overall system archite cture and interfaces and we also sketch the possibility of using Mathemati ca software in the client side of the system as well.\n\n10:45 - Coffee br eak\n\n11:00 - Oracle Servers Configuration Update - Eric Grancher - IT/DB Organiser(s): Pedro Martel / EST Division\n\nhttp://indico.cern.ch/confere nceDisplay.py?confId=a033146 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033146 END:VEVENT BEGIN:VEVENT SUMMARY:Anaphe - OO Libraries for Data Analysis using C++ and Python\n DTSTART;VALUE=DATE-TIME:20011205T150000Z DTEND;VALUE=DATE-TIME:20011205T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033147@cern.ch DESCRIPTION:The Anaphe project is an ongoing effort to provide an Object O riented software environment for future HENP experiments. A range of comme rcial and public domain libraries is used to cover basic functionalities\; on top of these libraries a set of HENP-specific C++ class libraries for histogram management\, fitting and ntuple-like data analysis has been deve loped. \nThe talk will give an overview of the architecture and design cho ices and will present the current status and future developments of the pr oject.Organiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/co nferenceDisplay.py?confId=a033147 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033147 END:VEVENT BEGIN:VEVENT SUMMARY:The Geant4 Toolkit Today DTSTART;VALUE=DATE-TIME:20011212T150000Z DTEND;VALUE=DATE-TIME:20011212T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033148@cern.ch DESCRIPTION:Geant4 is a detector simulation toolkit designed for the new g eneration of High Energy Physics experiments as well as for nuclear physic s\, medical and space applications\, radiation background studies. Approxi mately three years since the release of the first Geant4 production versio n\, significant results are already available from many applications of Ge ant4 in a variety of domains\, such as HEP experiments\, space and medical applications. New results are produced continuously. An overview of the s tatus of Geant4 and highlight results of its application will be presented . \nOrganiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/conf erenceDisplay.py?confId=a033148 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033148 END:VEVENT BEGIN:VEVENT SUMMARY:1. Linux Testing Framework
2.CERN Red Hat 7.2 Progress Report DTSTART;VALUE=DATE-TIME:20020116T150000Z DTEND;VALUE=DATE-TIME:20020116T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033149@cern.ch DESCRIPTION:1. Linux Testing Framework\n\nThe Testing Framework will allow you to test other programs in an automated way. Using the framework's fea tures allows you to more easily create powerful automated tests for your s oftware. The framework is primarily designed to run under the Linux operat ing system\, but has also been successfully tested under Sun Solaris. The seminar will cover the basics of the Testing Framework architecture\, how it works\, and its usage.\n\n2.CERN Red Hat 7.2 Progress Report\n\nAn upda ted version of Red Hat 7.2 Linux\, known as CERN Red Hat 7.2 Beta 1\, has been running in the CERN computer centre and on some workstations for a fe w weeks. The speaker will review the work done and outline problems solved so far and ones to be solved before a certified version of CERN Red Hat 7 .2 Linux can be released.\n\nOrganiser(s): Julian Blake / IT Division\n\nh ttp://indico.cern.ch/conferenceDisplay.py?confId=a033149 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033149 END:VEVENT BEGIN:VEVENT SUMMARY:Bitmap Indices for Speeding Up End User Physics Analysis DTSTART;VALUE=DATE-TIME:20020206T150000Z DTEND;VALUE=DATE-TIME:20020206T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033150@cern.ch DESCRIPTION:Bitmap indices are index structures that efficiently support c omplex\, multi-dimensional queries for datawarehouse applications. These d ata structures are also implemented in commercial database management syst ems such as Oracle\, Sybase or Informics. All these implementations are op timised for typical business applications which are characterised by discr ete attribute values. However\, scientific data which is mostly characteri sed by non-discrete attribute values\, cannot be handled efficiently by th ese kind of data structures.\nIn this seminar we give an overview of bitma p indices and demonstrate that this kind of access method shows good perfo rmance behaviour in multi-dimensional search spaces and significantly outp erforms the sequential scan for typical end user physics analysis.Organise r(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=a033150 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033150 END:VEVENT BEGIN:VEVENT SUMMARY:Access Networking DTSTART;VALUE=DATE-TIME:20020403T140000Z DTEND;VALUE=DATE-TIME:20020403T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033154@cern.ch DESCRIPTION:Rapid technological developments in the field of computers and telecommunications in the last 50 years have led to their convergence for information processing and transmission. Further\, during the last decade \, there has been reduction in the cost of equipments and transmission sys tems resulting in their wide spread availability and deployment. Moreover\ , there has been proliferation in the variety of service demands by the cu stomers in the recent years. As the core telecommunication networks mature with the variety of switching and transmission systems\, selection of app ropriate technology for accessing the core network by the end users for mu ltiple services at appropriate cost and QoS becomes a major issue. With th e predominance of EPABX/LANs at the customer premises\, access networks ar e becoming more demanding in terms of higher bandwidth to satisfy the appl ication needs. In addition\, traditional circuit switched based applicatio ns are migrating to packet switched technology for better channel utilizat ion. The selection of appropriate access technology becomes more complex d ue to variety in regulations across countries\, service providers offering new services and heterogeneous access channel environment (fixed\, mobile \, satellite etc) available to the user. Since no single technology is sui table for all the applications\, knowledge of various access technologies\ , QoS demands of various services\, cost and market trends are important b efore making long term investments. Panoramic view of these topics will be presented in the seminar.Organiser(s): Julian Blake / IT Division\n\nhttp ://indico.cern.ch/conferenceDisplay.py?confId=a033154 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033154 END:VEVENT BEGIN:VEVENT SUMMARY:J2EE or .NET ˆ an Objective Comparison DTSTART;VALUE=DATE-TIME:20020410T120000Z DTEND;VALUE=DATE-TIME:20020410T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032786@cern.ch DESCRIPTION:Microsoft and the open standards community? The conflict is an old one. As in any long running battle\, the ideology and history of form er conflicts can cloud the vision and approach of both sides on this great divide. Today the battle manifests itself in its most significant form ev er. NET or J2EE? Vendors and protagonists from both sides of the field pit ch their marketing spin and throw uncertainty and doubt in every direction . In this session I will try to take a completely objective and comparativ e view of the two specific technology stacks and to clearly present and de monstrate the issues and considerations of the two competing platforms. Th e session will be of real interest to anyone who wants to understand the r eal differentiators between J2EE and .NET. Whilst the session certainly le verages my skills and knowledge with the Oracle technology stack\, this is not an Oracle technology focused session. Rather I will attempt the impos sible and try to remain both objective and to guide your understanding to enable you to build the clarity and vision you need to comprehend how thes e two key technologies will shape the technology world in the coming years . \n\nBiography:\n\nDavid Keene is Senior Principal Product Manager within Oracle Server Technologies\, Oracle 9i Application Server\, EMEA. David h eads up the EMEA team of the engineering product management organisation f or Oracle 9iAS and leads the rollout of the new technologies into both the Oracle EMEA organisation and key strategic customers and partners. David has some 18 years in the IT industry from a start as a programmer through to his position today with some of the major consulting and financial orga nisations within the UK\, including Barclays\, American Express\, Royal Su n Alliance\, Lloyds of London\, Sema & Cap Gemini.\n\nhttp://indico.cern.c h/conferenceDisplay.py?confId=a032786 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032786 END:VEVENT BEGIN:VEVENT SUMMARY:Storage Futures DTSTART;VALUE=DATE-TIME:20020514T120000Z DTEND;VALUE=DATE-TIME:20020514T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033155@cern.ch DESCRIPTION:Given that today's HEP experiments generate enormous volumes o f data - several hundred TB of data per year in the case of COMPASS\, risi ng to several PB per year at the LHC - storage is one of the key areas tha t must be carefully tracked. This talk will address some of the main quest ions regarding the future of storage\, including \n\n- Versa Store in real ity: are there any products in development? \n- EVA next generation with a roadmap \n- DRM in Enterprise Virtual Array \n- SAN Management in future and the (possible) connection to OpenView \n- What kind of hardware exists in the lab today? What is planned? \n- Future SAN Architecture McData/Bro cade in the same SAN - OpenSAN. \n- Backup incl. server less backup/restor e. \n- NAS in a SAN \n- Disk caching machines \n- Infiniband \n- Disk vers us tape \n\nBIO\n\nStephen J. Sicola is Technical Director for Compaq\, St orageWorks. He has twenty-one years experience within the storage array bu siness. He currently holds 10 patents with 20 other patents pending in the areas of hardware\, software\, and packaging design.\nSteve graduated fro m Stanford University with a BS in Electrical Engineering '79\, and an MS in Computer Architecture '93.\n\nhttp://indico.cern.ch/conferenceDisplay.p y?confId=a033155 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033155 END:VEVENT BEGIN:VEVENT SUMMARY:J2EE or .NET ˆ an Objective Comparison DTSTART;VALUE=DATE-TIME:20020522T120000Z DTEND;VALUE=DATE-TIME:20020522T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032787@cern.ch DESCRIPTION:Microsoft and the open standards community? The conflict is an old one. As in any long running battle\, the ideology and history of form er conflicts can cloud the vision and approach of both sides on this great divide. Today the battle manifests itself in its most significant form ev er. NET or J2EE? Vendors and protagonists from both sides of the field pit ch their marketing spin and throw uncertainty and doubt in every direction . In this session I will try to take a completely objective and comparativ e view of the two specific technology stacks and to clearly present and de monstrate the issues and considerations of the two competing platforms. Th e session will be of real interest to anyone who wants to understand the r eal differentiators between J2EE and .NET. Whilst the session certainly le verages my skills and knowledge with the Oracle technology stack\, this is not an Oracle technology focused session. Rather I will attempt the impos sible and try to remain both objective and to guide your understanding to enable you to build the clarity and vision you need to comprehend how thes e two key technologies will shape the technology world in the coming years . \nBio:\nDavid Keene is Senior Principal Product Manager within Oracle Se rver Technologies\, Oracle 9i Application Server\, EMEA. David heads up th e EMEA team of the engineering product management organisation for Oracle 9iAS and leads the rollout of the new technologies into both the Oracle EM EA organisation and key strategic customers and partners. David has some 1 8 years in the IT industry from a start as a programmer through to his pos ition today with some of the major consulting and financial organisations within the UK\, including Barclays\, American Express\, Royal Sun Alliance \, Lloyds of London\, Sema & Cap Gemini.Note: This has been rescheduled (t he speaker was unable to make it due to air traffic problems resulting fro m a computer failure...) \n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a032787 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032787 END:VEVENT BEGIN:VEVENT SUMMARY:Data Recording Technology - past\, present & future DTSTART;VALUE=DATE-TIME:20020529T140000Z DTEND;VALUE=DATE-TIME:20020529T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033151@cern.ch DESCRIPTION:A driving force behind key decisions in data acquisition\, pro cessing\, sorting\, analyzing\, and storage of data is the quality of the recording system. Quality of design\, along with hardware\, recording medi a\, and the recording media interface can make a significant difference in the quality and reliability of resulting data. This talk will touch on va rious aspects of past and present data storage processes\, and look to the near future (3-5 years) to give personal views mixed with some known deve lopmental products that could shape the future of data storage. Hearing an other viewpoint on data storage is an opportunity to benefit and perhaps l earn some thing new about the importance of quality data recordings that c an compliment your decision process. Data can be considered 'priceless' wh en it is known that certain research and special event data can not be gen erated economically or collected a second time. This talk will offer some suggestions on keeping data readable after it is already in hand as well a s comments on improving data quality before recording.\n\n\n\nOrganiser(s) : Julian Blake / IT Division\n\nhttp://indico.cern.ch/conferenceDisplay.py ?confId=a033151 LOCATION:CERN 40-S2-B01 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033151 END:VEVENT BEGIN:VEVENT SUMMARY:The Internet Backplane Protocol and the L-Bone project proposal\n DTSTART;VALUE=DATE-TIME:20020530T090000Z DTEND;VALUE=DATE-TIME:20020530T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033152@cern.ch DESCRIPTION:This talk will explore the concepts and mechanisms underlying Logistical Networking\, a revolutionary architectural approach to communic ation that synthesizes elements of storage and wide area networking system s that are traditionally considered orthogonal. Logistical Networking is m odeled on IP networking\, and so its architecture is a stack with physical media and OS drivers at the bottom\; an innovative layer that enables the scalable sharing of storage called the Internet Backplane Protocol (servi ng a function analogous to IP)\; and then the exNode\, a tool for aggregat ing resources and enabling valuable end-to-end services such as reliabilit y\, high performance\, and security (serving a function analogous to TCP). Researchers at the Logistical Computing and Internetworking Laboratory at the University of Tennessee have been pursuing the development of these m echanisms and their integration into higher level middleware and\napplicat ion-level tools in an effort to create a new architecture for scalable com puting in the wide area based on the successful architectural approach of the Internet. This new architecture offers a new framework for advanced In ternet applications of all kinds\, from scientific Grid computing to colla borative work to multimedia content delivery. Of particular interest are c urrent efforts to deploy the Internet Backplane Protocol in a current mult i-terabyte testbed (the L-Bone\, mainly in the US) and a proposed petabyte -scale Logistical Networking Testbed that would span international univers ities and research labs.Organiser(s): Julian Blake / IT Division\n\nhttp:/ /indico.cern.ch/conferenceDisplay.py?confId=a033152 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033152 END:VEVENT BEGIN:VEVENT SUMMARY:Beating Moore's Law Through Software Techniques DTSTART;VALUE=DATE-TIME:20020605T123000Z DTEND;VALUE=DATE-TIME:20020605T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032788@cern.ch DESCRIPTION:At any given time\, parallelism and distributed computing can provide enormous performance gains for selected problems. The challenge is to provide performance benefits for the widest range of applications\, wh ile making it reliable and easy to use. Compiler technology plus developme nt tools and runtime support infrastructure are all needed for this\, and are improving. The talk will discuss some of the necessary software as wel l as underlying hardware necessities that Intel is developing to support s calable HPC.\n\nBIO:\nDavid J. Kuck is an Intel Fellow\, Enterprise Platfo rms Group and Director\, KAI Software Lab\, Software and Solutions Group. KAI is a leading provider of performance-oriented compilers and programmin g tools used in the development of multithreaded applications. Kuck founde d KAI in 1979 and was Chairman of the Board through 2000. Previously\, he was a faculty member of the Computer Science and Electrical and Computer E ngineering departments of the University of Illinois at Urbana-Champaign. He also served as director of the Center for Supercomputing Research and D evelopment. Kuck holds a B.S. in electrical engineering from the Universit y of Michigan\, as well as M.S. and Ph.D. degrees in engineering from Nort hwestern University. He is a fellow of the American Association for the Ad vancement of Science\, the Association for Computing Machinery\, and the I nstitute of Electrical and Electronics Engineers. He is also a member of t he National Academy of Engineering. Kuck holds two patents and has publish ed over 100 papers.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a 032788 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032788 END:VEVENT BEGIN:VEVENT SUMMARY:Performance Monitoring with AFFAIR DTSTART;VALUE=DATE-TIME:20020612T140000Z DTEND;VALUE=DATE-TIME:20020612T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033156@cern.ch DESCRIPTION:It is becoming increasingly important to be able to efficientl y and easily monitor the system performance of a very large computing fabr ic\, and also the applications running on it.\nI present in this talk AFFA IR\, the performance monitoring software developed by the ALICE Data Acqui sition project. AFFAIR is largely based on open source code and is compose d of the following components: data gathering\, inter-node communication e mploying DIM\, fast and temporary round robin database storage\, and perma nent storage and plot generation using ROOT. Real time data is monitored v ia a PHP generated web interface.\nAFFAIR is successfully used during the ALICE Data Challenge. It is scalable to hundreds of nodes\, robust\, easy to maintain\, flexible\, and has low network and CPU requirements.\n\nAbou t the speaker: \nTome Anticic earned his doctorate at the Johns Hopkins Un iversity in 1997\, working at the L3 experiment. For the past two years he has been involved in the ALICE DAQ team\, where he is leading the effort to develop a detailed simulation of the ALICE Data Acquisition and Trigger . He is also in charge of creating AFFAIR to monitor the ALICE Data Acquis ition System.\n\nOrganiser(s): Julian Blake / IT Division\n\nhttp://indico .cern.ch/conferenceDisplay.py?confId=a033156 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033156 END:VEVENT BEGIN:VEVENT SUMMARY:Interoperability and .NET DTSTART;VALUE=DATE-TIME:20020612T120000Z DTEND;VALUE=DATE-TIME:20020612T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033157@cern.ch DESCRIPTION:Over the last decade a number of Component Based Architectures (CBAs) have been developed to facilitate language interoperability. The m ost widely known and used of these would be the Object Management Group's (OMG's) Common Object Request Broker Architecture (CORBA) and Microsoft's Component Object Model (COM) and Distributed COM (DCOM). Although not spec ifically designed to provide language interoperability\, many Java based t echnologies\, such as Enterprise Java Beans (EJB)\, also adhere to the pri nciples of a CBA. Microsoft's .NET Framework is the latest CBA to be devel oped and builds on many of the lessons learnt from previous architectures. Interoperability is a core facility of the .NET Framework. The .NET frame work achieves language interoperability via the Common Language Runtime (C LR). The .NET framework also provides architecture interoperability via tw o mechanisms: COM Interoperability and Web Services. This talk firstly wil l give an architectural overview of the .NET Framework and explain how lan guage interoperability is achieved within the CLR. Then we will look at ho w architecture interoperability is achieved with both COM Interoperability and Web Services. Attendees will not need to be familiar with any CBA.\n\ nBio:\n\nDr Damien Watkins is the founder and Managing Director of Project 42\, a \ncompany that provides consultancy\, training\, mentoring and str ategic analysis on component-based software development for the Internet. Prior to commencing Project 42 Damien lectured at the School of Computer S cience and Software Engineering and the Department of Software Development at Monash University. Damien lectured both postgraduate and undergraduate subjects\, covering the areas of Component Based Software Development and Distributed Object Oriented Technologies. Damien taught the world's first .NET Framework course at Monash University . He has also taught at Uppsal a University (Sweden) and KMITNB (Thailand).\n\nhttp://indico.cern.ch/conf erenceDisplay.py?confId=a033157 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033157 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the IEEE/NASA Mass Storage Conference DTSTART;VALUE=DATE-TIME:20020626T140000Z DTEND;VALUE=DATE-TIME:20020626T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033153@cern.ch DESCRIPTION:The speaker will review the presentations given at the IEEE/NA SA mass storage conference (10th NASA-Goddard Conference on Mass Storage S ystems and Technologies / 19th IEEE Symposium on Mass Storage Systems\, Co llege Park\, Maryland\, 15-18 April 2002). The theme of this year's confer ence was to look at how current technological evolutions will impact the d evelopment and deployment of information storage and retrieval systems. \n About the speaker: \nFabien Collin is a member of IT/DS group. He has wide expertise in storage device management at all levels. Organiser(s): Julia n Blake / IT Division\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId =a033153 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033153 END:VEVENT BEGIN:VEVENT SUMMARY:SMP Parallelism: Current Achievements\, Future Challenges DTSTART;VALUE=DATE-TIME:20020703T120000Z DTEND;VALUE=DATE-TIME:20020703T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032790@cern.ch DESCRIPTION:SMP (symmetric Multi-Processors) hardware technologies are ver y popular with vendors and end-users alike for a number of reasons. Howeve r\, true shared memory parallelism has experienced somewhat slower take up amongst the scientific-programming community. NAG has been at the forefro nt of SMP technology for a number of years\, and the NAG SMP Library has s hown the potential of SMP systems. At the very high end\, SMP hardware tec hnologies are used as building blocks of modern supercomputers\, which tru ly consist of clusters of SMP systems\, for which no dedicated model of pa rallelism yet exists.\nThe aim of this talk is to introduce SMP systems an d their potential. Results from our work at NAG will also be introduced to show how SMP parallelism\, based on a shared memory paradigm\, can be use d to very good effect and can produce high performance\, scalable software . The talk also aims to discuss some aspects of the apparent slow take up of shared memory parallelism and the potential competition from PC (i.e. I ntel)-based cluster technology. The talk then aims to explore the potentia l of SMP technology within 'hybrid parallelism'\, i.e. mixed distributed a nd shared memory modes\, illustrating the point with some preliminary work carried out by the author and others. Finally\, a number of potential fut ure challenges to numerical analysts will be discussed. The talk is aimed at all who are interested in SMP technologies for numerical computing\, ir respective of any previous experience in the field. The talk aims to stimu late discussion\, by presenting some ideas\, backing these with data\, not to stifle it in an ocean of detail!\n\nBIO:\n\nDr Stefano Salvini is the Group Leader of the High Performance Computing Group of the Numerical Libr aries Division of NAG Ltd.\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=a032790 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032790 END:VEVENT BEGIN:VEVENT SUMMARY:Extreme Programming and Agile Methodologies: a new way to develop software systems DTSTART;VALUE=DATE-TIME:20020710T120000Z DTEND;VALUE=DATE-TIME:20020710T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032789@cern.ch DESCRIPTION:The seminar will present the Agile Approach to software develo pment\, an approach alternative to traditional software engineering\, whic h is gathering increasing interest among software developers. The principl es of the Agile Manifesto will be discussed\, and the main agile methodolo gies (XP\, Scrum\, Crystal\, FDD) will be briefly presented. Extreme Progr amming (XP)\, which is the most popular among AM will then be discussed mo re in detail\, highlighting the pros and cons of its use with respect to t raditional software development processes. \n\nBio:\n\nMichele Marchesi is a professor at the University of Cagliari\, Italy\, author of a number of books on Extreme Programming and Program Chair for the XP 200n conference series.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a032789 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032789 END:VEVENT BEGIN:VEVENT SUMMARY:Distributed Application Analysis and Debugging using NetLogger2 Grid Programming in Python using pyGlobus DTSTART;VALUE=DATE-TIME:20020716T140000Z DTEND;VALUE=DATE-TIME:20020716T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033161@cern.ch DESCRIPTION:Netlogger (short for Networked Application Logger) is a method ology for performing detailed end-to-end analysis of distributed applicati ons. It includes tools for instrumenting applications\, host systems\, and networks\, and has a powerful visualization tool for correlating monitori ng data from all components. NetLogger has proven to be extremely useful f or debugging and tuning distributed applications. We have recently release d NetLogger2\, which has a number of important enhancements\, including a new efficient binary event format\, an automatic recovery mechanism\, and a web-service-based activation mechanism. This talk will present a brief o verview of NetLogger\, and then concentrate on new NetLogger2 features. Th e talk will also cover how we are using NetLogger2 as the data transport f or our implementation of the Grid Monitoring Architecture (GMA).\npyGlobus provides a high-level object-oriented interface to the Globus Toolkit(tm) and allows the rapid development of Grid applications. It provides an int erface to all of the major components of the Globus Toolkit\, including se cure IO\, resource acquisition and control\, data transfers\, etc. pyGlobu s has been used to implement a number of Grid applications\, and has prove n very useful for decreasing both time to delivery and cost of application development. This talk will provide an overview of the current pyGlobus f unctionality\, including our recent work on secure web services. It will a lso include an overview of our current plans for a Python based Open Grid Services Infrastructure (OGSI) implementation.Organiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033161 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033161 END:VEVENT BEGIN:VEVENT SUMMARY:Scalable Data Access in Peer-to-Peer Systems DTSTART;VALUE=DATE-TIME:20020731T140000Z DTEND;VALUE=DATE-TIME:20020731T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033162@cern.ch DESCRIPTION:With the appearance of Peer-to-Peer information systems the in terest in scalable and decentralized data access structures is attracting increasingly interest. We propose to that end P-Grid\, a scalable data acc ess structure resulting from the distribution of a binary prefix tree. Whe n adapting the P-Grid structure to skewed data distributions one obtains u nbalanced search trees. We show that unbalanced trees do not harm as long as communication is considered as the critical cost and the access structu res are constructed properly. We propose the necessary distributed\, rando mized algorithms that allow to construct the P-Grid in a self-organized ma nner such that the tree structure dynamically adapts to the data distribut ion and the aforementioned result is applicable. We compare our approach t o other distributed indexing schemes that have been proposed in research a nd practice.Organiser(s): Peter Kunszt / IT Division\n\nhttp://indico.cern .ch/conferenceDisplay.py?confId=a033162 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033162 END:VEVENT BEGIN:VEVENT SUMMARY:Research in Storage Systems at IBM Research DTSTART;VALUE=DATE-TIME:20020807T120000Z DTEND;VALUE=DATE-TIME:20020807T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032791@cern.ch DESCRIPTION:Jai Menon will discuss Research in Storage Systems at IBM Rese arch. He will describe work in scalable\, global file systems\, storage vi rtualization\, storage QOS\, autonomic storage systems and Grid storage. P rojects that will be covered will include Storage Tank and Icecube. Storag e Tank is an enterprise-wide\, geographically distributed\, storage softwa re infrastructure which permits goal oriented\, policy-driven storage mana gement. IceCube is an IBM Research-designed architecture for servers that follows these simple principles. It consists of a large\, three-dimensiona l array (the 'Cube') of 'bricks' that collectively provide a very high res ilience against failure. In a storage server application\, each brick cont ains multiple disks and a processor. An extremely high bandwidth\, three-d imensional mesh connects the bricks\, making the location of data within t he cube nearly irrelevant. This makes it possible for software to scatter and replicate data over many independent bricks. The scheme provides high assurance against data loss\, as well as high performance for large data s ets because subsets of the data can be retrieved in parallel.\n \nBIO:\nDr . Jai Menon is IBM Research's leading authority on the design and architec ture of data storage systems\, which enable enterprises and organizations to manage their data reliably and efficiently. A native of Kerala\, India\ , Menon received a Bachelor of Technology degree in Electrical Engineering from the Indian Institute of Technology\, Madras\, India in 1977. He earn ed a Master of Science and Ph.D degrees in Computer Science from Ohio Stat e University in 1978 and 1981\, respectively. In 1982\, Menon joined IBM R esearch in San Jose\, Calif.\, where he became a pioneering researcher and designer of data storage systems and RAID (Redundant Array of Independent Disks) architectures. He was named functional manager\, CS\; Storage Syst ems\, in April 2000.\nIn May 2001\, Jai was named IBM Fellow\, the company 's most prestigious and highest technical honor. He is also co-director of the Storage Systems Institute\, a virtual organization/joint program betw een IBM Research and the Storage Systems Group aimed at speeding the incor poration of scientific and technological advances into IBM storage systems products. An IEEE Fellow and member of the IBM Academy of Technology\, Me non has received numerous IBM technical awards\, \npublished 31 refereed p apers and 46 technical reports and is contributing author to three books o n database and storage systems. He also holds 42 U.S. Patents and has 20 a dditional patent applications on file with the U.S. Patent and Trademark O ffice.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a032791 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032791 END:VEVENT BEGIN:VEVENT SUMMARY:eScience\, eBusiness and the Grid DTSTART;VALUE=DATE-TIME:20020821T123000Z DTEND;VALUE=DATE-TIME:20020821T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032792@cern.ch DESCRIPTION:After defining what is meant by the term 'e-Science'\, this ta lk will survey the activity on e-Science and Grids in Europe. The two larg est initiatives in Europe are the European Commission's portfolio of Grid projects and the UK e-Science program. The EU under its R&D Framework Prog ram are funding nearly twenty Grid projects in a wide \nvariety of applica tion areas. These projects are in varying stages of maturity and this talk will focus on a subset that have most significant progress. These include the EU DataGrid project led by CERN and two projects 'EuroGrid and Grip' that evolved from the German national Unicore project. A summary of the ot her EU Grid projects will be included. The UK e-Science initiative is a $1 80M programme entirely focused on e-Science applications requiring resourc e sharing\, a virtual organization and a Grid infrastructure.\n\nThe talk will conclude with a survey of other Grid initiatives across Europe and lo ok at possible future European \nprojects.\n\nNote: Tony Hey is the Direct or of e-Science for The Engineering and Physical Sciences Research Council (EPSRC)\, which is \nthe UK's leading funding agency for research and tra ining in engineering and the physical sciences.\n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=a032792 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032792 END:VEVENT BEGIN:VEVENT SUMMARY:Logistical Networking: A Scalable Storage Architecture for Data In tensive Collaboration DTSTART;VALUE=DATE-TIME:20020909T120000Z DTEND;VALUE=DATE-TIME:20020909T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033160@cern.ch DESCRIPTION:The current model for collaboration in the wide area is to bui ld isolated islands of resources seperately owned and controlled by collab orating organizations and to connect them using data transmission networks \, generally IP-based. In this talk\, we will propose Logistical Networkin g as a new model in which organizations provision storage resources for th e purpose of sharing them with their collaborators\, much as data transmis sion resources are provisioned in an IP network. The premise of this propo sal is that in a data intensive collaborative environment\, it is importan t for each collaborator to be able to flexibly utilize storage resources a t any location in the network. One of the reasons that IP networking has b ecome the ubiquitous mechanism for sharing data transmisison resources has been its scalability\, a characteristic that is largely due to observing the so-called 'end-to-end' architectural principles. One of the reasons th at storage resources could not in the past be widely deployed for shared u se is that there was no network storage mechanism that similarly observed those principles. We will introduce a new storage mechanism\, the Internet Backplane Protocol (IBP)\, that is the foundation of the Logistical Netwo rking architecture and show that it does observe the end-to-end principles . We will also describe the other components of the Logistical Networking architecture\, which include resource discovery and 'end-to-end' services that ensure reliability\, high performance and security. Researchers at th e Logistical Computing and Internetworking Laboratory at the University of Tennessee have been pursuing the development of these mechanisms and thei r integration into middleware and applications. Data intensive science app lications are of particular interest\, as they place a uniquely heavy load on data transmission resources. Of particular interest is the possibility of integrating high bandwidth data transmission mechanisms other than TCP /IP to transport data between IBP storage nodes while still maintaining in teroperability with current IP research networks. Logistical Networking de ployment efforts are focused on a multi-terabyte Logsitical Backbone testb ed\, currently in the U.S. and Europe\, and a proposed petabyte-scale Logi stical Networking Testbed that would span universities and research labs w orldwide.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033160 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033160 END:VEVENT BEGIN:VEVENT SUMMARY:DIANE Distributed Analysis Environment DTSTART;VALUE=DATE-TIME:20020911T140000Z DTEND;VALUE=DATE-TIME:20020911T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033158@cern.ch DESCRIPTION:DIANE is a CERN IT/API project to study the requirements and p rototype a middleware distributed environment for parallel data analysis f or LHC. The main goal of DIANE is to provide LHC experiments with applicat ion-oriented component-based framework for parallel data analysis. DIANE i s targeted to handle mainly parallel ntuple analysis and parallel detector simulation but is not a priori technically limited to any specific applic ation\, and thus may be used in a wider context. DIANE is independent from any particular data analysis software and is explicitly designed in such a way that analysis-specific parts are covered as a separate component. Th is fits well with the LCG project architecture and makes the future LCG ap plication components very easy to use in the context of DIANE. The baselin e architecture of DIANE uses CORBA 2.3 and CCM\, an industry standard for distributed computing and component technology. DIANE uses several GRID co mponents for low-level services such as environment replication\, security \, load balancing and the like. From a top-down perspective\, a cluster of machines running DIANE is the GRID computing element and may be accessed via GRID-enabled web portals and directly via the GRID API. DIANE may also be used directly in interactive analysis tools and may be embedded in a p ython environment.\nOrganiser(s): Julian Blake / IT Division\nhttp://wwwin fo.cern.ch/seminars/\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId= a033158 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033158 END:VEVENT BEGIN:VEVENT SUMMARY:Inductive Logic Programming: Towards machine assisted theory forma tion? \n DTSTART;VALUE=DATE-TIME:20020911T120000Z DTEND;VALUE=DATE-TIME:20020911T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033159@cern.ch DESCRIPTION:How can a machine assistant help in the design of new drugs? W hy do proteins fold one way and not another? Who causes accidents on the L ondon orbital motorway? What strategies are hackers using to break into th e system? How can organisations recognise 'High Potential' managers? What rules did Rachmaninoff use to play his piano? Inductive Logic Programming (ILP) -- a research area formed at the intersection of Machine Learning an d Logic Programming -- is being applied to construct answers to these\, an d many other questions. ILP systems construct predicate descriptions from examples and background knowledge. The examples\, background knowledge and final descriptions are all described as logic programs. A unifying theory of Inductive Logic Programming is being constructed from lattice-based co ncepts such as refinement\, least general generalisation\, inverse resolut ion and most specific corrections. Presently successful applications areas for ILP systems include problems in science\, engineering\, language and music. Some results from ILP systems are already published scientific disc overies. This talk will provide a short introduction to the curious world of ILP. It is aimed towards the application of ILP\, and only presents bas ic theoretical ILP concepts.\n\nhttp://indico.cern.ch/conferenceDisplay.py ?confId=a033159 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033159 END:VEVENT BEGIN:VEVENT SUMMARY:SST-1 Data Acquisition and Control System DTSTART;VALUE=DATE-TIME:20020918T140000Z DTEND;VALUE=DATE-TIME:20020918T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033163@cern.ch DESCRIPTION:SST-1 is a steady state Tokamak device which is designed for f undamental study of plasma physics. SST-1\, being a steady state device\, will require real time control and data handling\, analysis and displaying capabilities. The continuous nature of operation has a significant effect on instrumentation and control\; the internal architecture of digitiser a nd controller also plays a major role on performance. \n\nSince shot durat ion is 1000 seconds\, to limit the volume of data generated different stra tegies have been adopted for slow and fast diagnostics. The low frequency diagnostics demanding continuous acquisition will be based on PXI system d irectly streaming to hard disk and network for online storage and viewing. High frequency diagnostic data will be acquired on known events and will be stored in different segments of memory. Special purpose CAMAC modules h ave been developed in house.\n\nSST control system will be based on a dist ributed control system. For efficient management of system components\, th ey have been divided into three groups: Machine Control System\, Discharge Control System and Diagnostics Supervisory Control. Machine control syste m will configure and operate various technical subsystems with a hierarchi cal control structure.\n\nThe discharge control system will be a real time control system based on VME and RTOS which will take decisions regarding operational phases and will control plasma parameters. Discharge control w ill be active during discharge and discharge related subsystems would be u nder direct control of discharge control supervisor. The system will be us ing reflective memory for real time digital data transfers across the vari ous subsystems.Organiser(s): Julian Blake / IT Division\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=a033163 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033163 END:VEVENT BEGIN:VEVENT SUMMARY:Spring 2002 CMS Monte Carlo Production: what? how? and what next?\ n DTSTART;VALUE=DATE-TIME:20020925T140000Z DTEND;VALUE=DATE-TIME:20020925T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033164@cern.ch DESCRIPTION:During the first four months of 2002\, the CMS Monte Carlo Pro duction team has produced about 20 TB of data for the preparation of the D AQ TDR [data acquisition technical design report]. It included the simulat ion of primary proton-proton interactions\, secondary interactions in the detector\, response of the sensitive elements\, pile-up of events\, and fi rst-pass analysis. The data was written into Objectivity/DB ODBMS. This wa s a world-wide production involving more than 30 persons and 20 sites. Thi s seminar presents the full machinery of this successful Monte Carlo produ ction and describes what worked and what didn't. Some first thoughts will also be presented on how to improve the system for the coming major produc tion preceding the 2004 data challenge.Organiser(s): Julian Blake / IT Div ision\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033164 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033164 END:VEVENT BEGIN:VEVENT SUMMARY:Grid Economics: Grid Value and Practical Realization DTSTART;VALUE=DATE-TIME:20020927T123000Z DTEND;VALUE=DATE-TIME:20020927T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032793@cern.ch DESCRIPTION:Grid computing\, for both processing and data\, has attracted great interest for its potential ability to increase the value of existing and future infrastructure. This increase in value has two main axes\, gre ater efficiency and greater flexibility. We will present a theoretical bas is for Grid computing value and describe the difference between what is th eoretically present and what can be practically attained. As might be expe cted full value realization requires changes in how resources are managed and appropriate tools.\n\nBIO:\n\nChris Kenyon is a Research Staff Member at IBM Zurich Research Laboratory. He has a MCSE in Operations Research fr om Univ. of Texas at Austin and a Ph.D. from Cambridge University. He has experience in commodity markets and price dynamics for both conventional c ommodities (e.g. oil) and also with more recent developments (e.g. bandwid th) as well as pricing of non-traded assets relative to market prices (rea l options). Current research concentrates on Internet economics especially Grid Economics. He is a member of INFORMS\, IEEE\, and BFS.\n\nhttp://ind ico.cern.ch/conferenceDisplay.py?confId=a032793 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032793 END:VEVENT BEGIN:VEVENT SUMMARY:Putting the power of digital communication into pen and paper DTSTART;VALUE=DATE-TIME:20021009T120000Z DTEND;VALUE=DATE-TIME:20021009T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032794@cern.ch DESCRIPTION:Pen and paper is one of the biggest information technologies i n the world\, and it is still to be connected to the digital domain. Anoto has invented a system combining a digital pen with application-specific p aper. In the talk we will explain how paper becomes digital screens with a completely natural user experience\, and how the Anoto infrastructure con nects any given paper with any service across the Internet.\n\nBIOS:\n\nÖ rjan Johansson\, Chairman of the Board\, Anoto Group AB is the dynamic for ce behind the creation of a broad coalition of partners supporting Anoto's standard for digital paper and digital pens enabling Anoto functionality. Before joining Anoto\, Johansson spearheaded the global development of Bl uetooth -the short-range radio technology\, as Director \nand General Mana ger for Bluetooth at Ericsson Mobile Communications AB. For nearly two yea rs Johansson was responsible for developing the technology and leading oth er competing companies to adopt the open standard. Johansson initiated and supervised the establishment of the Bluetooth Special Interest Group\, ma de up of adopter companies\, and was responsible for bringing 2000 interna tional technology companies on board. Mr. Johansson is a seasoned veteran of international business and technology with over twenty-years managing m ultinational teams and \nprojects working in and around technological inno vation. Örjan Johansson currently resides in Lund\, Sweden. He has a Mast er's of Science in Electrical Engineering from the Lund Institute of Techn ology.\n\nPetter Ericson is one of the principal inventors of the Anoto Te chnology. He joined Anoto's parent company C Technologies at the very star t\, and was managing the software development team when the Anoto concept was developed. Mr Ericson is now Chief Science Officer at Anoto\, with foc us on strategic technology and business development. His speciality at the MSc. in Engineering Physics programme at Lund Technical University was co mputational physics. He also studied at Santa Clara University for a schol arship year. Petter Ericson currently resides in Malmö\, Sweden.\n\nhttp: //indico.cern.ch/conferenceDisplay.py?confId=a032794 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032794 END:VEVENT BEGIN:VEVENT SUMMARY:Peer-to-Peer Grid Databases for Service and Resource Discovery DTSTART;VALUE=DATE-TIME:20021009T140000Z DTEND;VALUE=DATE-TIME:20021009T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033165@cern.ch DESCRIPTION:The convergence of Grid Computing\, Peer-to-Peer Computing\, D istributed Databases and Web Services promises powerful emerging synergies . Grids are collaborative distributed Internet systems characterized by la rge scale\, heterogeneity\, multiple autonomous administrative domains\, u nreliable components and frequent dynamic change. Here\, it is desirable t o maintain and query dynamic and timely information about active participa nts such as services\, resources and user communities. The web services vi sion promises that programs are made more flexible\, adaptive and powerful by querying Internet databases (registries) at run time in order to disco ver information and network attached building blocks\, enabling the assemb ly of distributed higher-level components. However\, in a large cross-orga nizational system\, the set of information tuples is partitioned over many such distributed databases\, for reasons including autonomy\, scalability \, availability\, performance and security. This suggests the use of Peer- to-Peer (P2P) query technology. We describe a P2P framework and network pr otocol that allow to express specific discovery applications for a wide ra nge of data types\, node topologies (e.g. ring\, tree\, graph)\, query lan guages (e.g. XQuery\, SQL)\, query response modes\, neighbour selection po licies\, pipelining\, timeout and scope policies.Organiser(s): Julian Blak e / IT Division\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a0331 65 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033165 END:VEVENT BEGIN:VEVENT SUMMARY:Tutorial: Software Packaging with RPM\n DTSTART;VALUE=DATE-TIME:20021023T140000Z DTEND;VALUE=DATE-TIME:20021023T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033166@cern.ch DESCRIPTION:RPM\, the Red Hat Package Manager tool\, has established itsel f as the software distribution format of choice for most major Linux distr ibutions. It is also the primary distribution format used in the European DataGrid project\, and possibly in LCG. This tutorial presents motivation\ , concepts and best practices of software packaging using RPM. It is inten ded to help software developers at CERN with the transition from other met hods of delivery (over AFS or as .tar files) toward RPM\, using real-life examples.Organiser(s): Julian Blake / IT Division\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=a033166 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033166 END:VEVENT BEGIN:VEVENT SUMMARY:Evaluation of the C++ Binding to the Oracle Database System\n DTSTART;VALUE=DATE-TIME:20021120T150000Z DTEND;VALUE=DATE-TIME:20021120T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033167@cern.ch DESCRIPTION:With version 9i of the Oracle data management system the Oracl e C++ Call Interface\, OCCI\, was introduced. CERN's Information Technolog y Database Group was actively involved in the product's early beta testing program. This seminar will give a basic overview of OCCI features and rep ort on the evaluation of using an object-relational database management sy stem together with its C++ binding in the context of high-energy physics' data storage and applications.Organiser(s): Julian Blake / IT Division\n\n http://indico.cern.ch/conferenceDisplay.py?confId=a033167 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033167 END:VEVENT BEGIN:VEVENT SUMMARY:Service Oriented Architectures & Grid Computing DTSTART;VALUE=DATE-TIME:20030115T130000Z DTEND;VALUE=DATE-TIME:20030115T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033084@cern.ch DESCRIPTION:The benefits of Web services will catalyze a trend towards ser vice-oriented architectures in which enterprises view their systems as an orchestrated Web of services that extend beyond their firewalls into the s ystems of their partners and customers. Service-oriented architectures req uire a new kind of distributed computing platform that provides features l ike clustering\, load balancing and fail over in a way that is independent of the services themselves.\nService-oriented grid-computing platforms ta ckle this challenge by operating in a similar manner to the national elect ricity grid. Such platforms connect producers and consumers of services an d data\, while shielding them from details of the underlying infrastructur e. This course discusses the technical challenges of implementing service- oriented grid computing platforms\, with a focus on P2P architectures.\nBI OGRAPHY\nGraham Glass is the CEO of The Mind Electric. Prior to founding T he Mind Electric\, Graham was the CTO and co-founder of ObjectSpace\, a Da llas-based company specializing in business-to-business integration. He re ceived an Ernst and Young Entrepreneur of the year award in 1996\, as well as several industry awards for the development of the Voyager and JGL pro duct lines while with ObjectSpace. Graham is the author of Prentice Hall's flagship book on web services titled "Web Services: Building Blocks for D istributed Systems". As a leader and visionary on the future of distribute d computing\, Graham has spoken at many industry conferences and trade sho ws. Graham earned his BSc in Mathematics and Computer Science from the Uni versity of Southampton\, and his MS in Computer Science from the Universit y of Texas at Dallas. In 1980\, he was awarded the British Computing Socie ty prize for the top computer science student in the South of England.\n\n http://indico.cern.ch/conferenceDisplay.py?confId=a033084 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033084 END:VEVENT BEGIN:VEVENT SUMMARY:Open Grid Service Architecture Data Access and Integration Project Overview \n DTSTART;VALUE=DATE-TIME:20030123T130000Z DTEND;VALUE=DATE-TIME:20030123T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033085@cern.ch DESCRIPTION:The UK based Open Grid Service Architecture - Data Access and Integration (OGSA-DAI) project is currently designing and building a middl eware reference implementation of the components required to access and in tegrate data in a Grid environment. The objective is that Grid data servic es will provide the ability to discover\, manage\, and access data resourc es within an OGSA compliant architecture. \nThe OGSA-DAI reference impleme ntation supports access to data in relational databases (MySQL\, DB2 and O racle)\, and data in XML repositories (Xindice). This presentation provide s an overview of the OGSA-DAI architecture\, together with Grid data servi ce functional capabilities and future direction.\nBIOGRAPHY\nDave Pearson started his career in IT more than 25 years ago after initially working as an exploration geologist. He spent a number of years in the oil industry developing grid-like applications which integrated data from different sou rces and visualised the results to support exploration and production proc esses. For the past 14 years he has worked for Oracle Corporation UK. He u ndertook consulting assignments in all industrial sectors before moving in to management. He has been involved in Grid activities for the past year a nd is a member of the UK Grid Database Taskforce. \n\nhttp://indico.cern.c h/conferenceDisplay.py?confId=a033085 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033085 END:VEVENT BEGIN:VEVENT SUMMARY:The GNU Scientific Library DTSTART;VALUE=DATE-TIME:20030129T150000Z DTEND;VALUE=DATE-TIME:20030129T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033086@cern.ch DESCRIPTION:The GNU Scientific Library (GSL) is a free numerical library f or the GNU operating system. The goal of the project is to provide a free equivalent to existing proprietary numerical libraries\, such as NAG or I MSL. This talk will give an introduction to the design and use of the lib rary\, and the history of GSL.Organizer: Julian Blake / IT Division\nInfor mation: http://wwwinfo.cern.ch/seminars/\n\nhttp://indico.cern.ch/conferen ceDisplay.py?confId=a033086 LOCATION:CERN PS Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033086 END:VEVENT BEGIN:VEVENT SUMMARY:Oracle Backup & Recovery DTSTART;VALUE=DATE-TIME:20030204T073000Z DTEND;VALUE=DATE-TIME:20030204T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033087@cern.ch DESCRIPTION:The Swiss Oracle User Group (SOUG) will offer on Tuesday Febru ary 4th several presentations about "Oracle Backup & Recovery".\nVarious t opics will be addressed:\n- Backup & recovery at CERN (C. Delamare - CERN\ , IT Division)\n- Backup & recovery with the Oracle RMAN tool (R. Maag - T rivadis SA\, Lausanne)\n- What's new in Oracle 9i ? (D. Verdier - Oracle S witzerland\, Genève)\nThe inscription is free of charge for CERN members\ , nevertheless it is necessary to register. If you wish to attend the lunc h that will follow\nthe seminar\, please register too (lunch price is 30CH F).\nFor any information and/or registration\, please contact Stephan Peti t\, EST Division (stephan.petit@cern.ch phone: 74765)Organiser(s): Stepha n Petit\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033087 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033087 END:VEVENT BEGIN:VEVENT SUMMARY:High Performance Network Storage DTSTART;VALUE=DATE-TIME:20030225T130000Z DTEND;VALUE=DATE-TIME:20030225T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032784@cern.ch DESCRIPTION:DataDirect Networks' Silicon Storage Appliance (S2A) is a fami ly of intelligent storage controllers which provides servers\, workstation clusters and standalone computers in a heterogeneous environment with eas y\, scalable and high bandwidth access to shared data. With built-in hardw are RAID\, a massive cache\, storage and LUN virtualization\, parallel Fib re Channel port technology\, and support for up to hundreds of terabytes o f storage\, DataDirect's S2As provide cost-effective and reliable support for the high bandwidth and large storage requirements of demanding high pe rformance computing and rich media networks.\nThe company's SwiftCluster f ile system uses the open-source Lustre file system\, based on Linux and de veloped by Lawrence Livermore\, to provide all 1\,000 users at the lab wit h a single view into their storage. DataDirect will supply Lawrence Liverm ore with 16 of its SAN appliances\, totaling 300 Tbytes\, to tie into 64 L inux boxes\, creating the world's fastest and cheapest Linux cluster.\nSwi ftCluster is a technological breakthrough which combines DataDirect's S2A8 000 parallel block level storage controller\, its scalable serial PCI I/O node interface\, Direct Memory Access optimized protocol layer\, all tying into a highly efficient parallel file system\, to create a blueprint for the future of clustering. With SwiftCluster\, customers will be able to bu ild very efficient computer clusters scaling to tens of teraflops\, while benefiting from Object Based Disk (OBD) file system technology.\nThe modul arity of the S2A8000 enables various front access interfaces\, Fibre Chann el\, Serial PCI\, Infiniband and iSCSI while supporting high performance ( over GB's/s per unit) storage access to very large centralized pool of Fib re or SATA drives (100's of TB's to PB's).\nDataDirect has designed\, depl oyed and supported world class storage networks for customers such as Lawr ence Livermore\, NASA\, Sandia\, NCSA\, Time Warner\, Sony and Pinnacle. T hese systems range in performance from less than 1GB/sec to 10GB/sec+\, in capacity from less than 5TB to 200TB+\, and support heterogeneous operati ng systems\, file systems and compute platforms.\nBIOGRAPHY\nMr. Paul Bloc h\, President\, Co-Founder Paul Bloch\, co-founder of DataDirect Networks\ , brings over 15 years of experience with start-up and management business es including serving as the president of Personal Writer\, Inc.\, and Mega Drive Systems\, Inc. Previously\, Mr. Bloch enjoyed a successful tenure as a director for Hi-Tech Consulting Group\, a European management consultin g firm based in the United States. Mr. Bloch graduated with an Engineering degree from National Superior School of Telecommunications (ENST) in Fran ce and a Masters in Electrical Engineering from the California Institute o f Technology (Caltech).\n\nhttp://indico.cern.ch/conferenceDisplay.py?conf Id=a032784 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032784 END:VEVENT BEGIN:VEVENT SUMMARY:The HippoDraw analysis application and the library upon which it i s built DTSTART;VALUE=DATE-TIME:20030312T140000Z DTEND;VALUE=DATE-TIME:20030312T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033088@cern.ch DESCRIPTION:HippoDraw is a highly interactive data analysis application wi th many features not found in other Open Source or commercial data analysi s applications. HippoDraw can be used as a stand-alone application\, as a module in an interactive Python session\, or as a module in a PyQt applic ation. In each case\, all the highly desirable interactive features of Hi ppoDraw are available. The C++ library upon which HippoDraw is built is v ery modular and easily extendable and/or useable in special applications.O rganiser(s): Julian Blake / IT Division\nInformation: http://wwwinfo.cern. ch/seminars/\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033088 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033088 END:VEVENT BEGIN:VEVENT SUMMARY:Mathematical Services and the Semantic Web DTSTART;VALUE=DATE-TIME:20030410T120000Z DTEND;VALUE=DATE-TIME:20030410T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a032785@cern.ch DESCRIPTION:Web and grid services are being promoted by IBM\, Microsoft an d others as a revolutionary new way to deliver applications to users. At the same time the Worldwide Web Consortium is developing the infrastructur e for a new semantic web where the meaning of information is made explicit so that it can be accessed and manipulated by computers as well as human beings. This talk will look at ways in which semantic web technology is b eing used inside the European Union's OpenMath and MONET projects to descr ibe what functions a mathematical web service performs and\, using that in formation\, select the most appropriate service to solve a particular prob lem.Biography\nMike Dewar is the leader of the XML Technologies Group at N AG Ltd based in Oxford\, and also manages the MONET and OpenMath-TN Europe an Fifth Framework projects. He has a PhD from the University of Bath in the are a of computer algebra and worked for many years on the Axiom and R educe systems.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a03278 5 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a032785 END:VEVENT BEGIN:VEVENT SUMMARY:InfiniBand Architecture in High Performance Computing Clusters DTSTART;VALUE=DATE-TIME:20030506T120000Z DTEND;VALUE=DATE-TIME:20030506T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033089@cern.ch DESCRIPTION:The InfiniBand Architecture (IBA) is a new interconnect techno logy whose purpose is to provide a scalable switched fabric for both serve r-to-server and server-to-I/O communication. In addition to low latency a nd high bandwidth\, IBA supports many new mechanisms and services\, such a s multiple transport services\, send-recv & memory-to-memory semantics\, a nd prioritised data flows. This presentation will provide an in-depth look at the InfiniBand Architecture. In addition\, features specific to the h igh performance computing community will be covered\, such as current MPI implementations over IB\, as well as the latest benchmarking results compa ring InfiniBand\, Myrinet and Quadrics interconnect technologies.Biography \nBrad Benton is senior architect for the InfiniSwitch Corporation's enter prise fabric management software products. \nBenton has more than 20 years of industry experience\, including eight in architecture positions. Most recently Benton served as Senior Technical Relationship Manager at Lane15 Software. Prior to this Benton was Development Manager at Infoglide Softwa re where he was responsible for defining technical direction\, product def inition\, and system performance. Before Infoglide he held senior position s at Pervasive Software and Tandem Computers.\n\nhttp://indico.cern.ch/con ferenceDisplay.py?confId=a033089 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033089 END:VEVENT BEGIN:VEVENT SUMMARY:Colloquium: Accurate High Performance Maths Libraries and Their Te sting DTSTART;VALUE=DATE-TIME:20030521T120000Z DTEND;VALUE=DATE-TIME:20030521T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033090@cern.ch DESCRIPTION:When building numerical libraries for modern high-performance architectures\, it is important to ensure that as well as creating fast co de we do not lose sight of the fact that the answers must be correct. With particular reference to recent work in collaboration with chip manufactur ers AMD for their new Opteron 64-bit processors\, we discuss the ways and means by which we try to assure correctness\, while giving some insight in to NAG's methods of quality control. We will look at low level routines su ch as trig functions from the standard mathematical library libm\, as well as higher level numerical codes such as BLAS and LAPACK\, and also show s ome comparative graphs of timing results.Biography\nMick Pont has been wit h NAG since 1984\, spending time in NAG's UK and North American offices. H e is currently a senior member of NAG's Development Division\, in Oxford. During his time at NAG Mick has worked on various topics including testing of floating-point arithmetic\, numerical library development\, and the vi sualisation system IRIS Explorer.\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=a033090 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033090 END:VEVENT BEGIN:VEVENT SUMMARY:Dynamic Parallel Schedules\, a Framework for Developing Parallel A pplications DTSTART;VALUE=DATE-TIME:20030702T120000Z DTEND;VALUE=DATE-TIME:20030702T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033091@cern.ch DESCRIPTION:Dynamic Parallel Schedules (DPS) is a high-level framework for developing parallel applications on distributed memory computers (e.g. cl usters of PCs or workstations). Its model relies on compositional customiz able split-compute-merge graphs of operations (directed acyclic flow graph s). The graphs and the mapping of operations to processing nodes are speci fied dynamically at runtime. DPS applications are pipelined and multithrea ded by construction\, ensuring a maximal overlap of computations and commu nications. DPS applications can call parallel services exposed by other DP S applications\, enabling the creation of reusable parallel components. Th e DPS framework relies on a cross-platform C++ class library\, allowing DP S applications and services to run on heterogeneous clusters. Thanks to it s dynamic nature\, DPS offers new perspectives for the creation and deploy ment of parallel applications running on server clusters.Short biographies Sebastian Gerlach is a research assistant and PhD candidate at the Ecole P olytechnique Fédérale de Lausanne (EPFL). He received a diploma in micro engineering from the EPFL. His current research interests include real-tim e 3D visualization and parallel application frameworks.Roger D. Hersch is professor of computer science and head of the Peripheral Systems Laborato ry at the Ecole Polytechnique Fédérale de Lausanne. He received his eng ineering and PhD degrees respectively from ETH Zurich in 1975 and from EPF L in 1985. He has published over 100 scientific papers\, is the editor of 4 books and is inventor or co-inventor in several patent applications. Th e labs research is focussed on high-performance server applications (imagi ng servers\, Web servers\, PC clusters\, etc...) and novel imaging techniq ues (color prediction\, color reproduction\, artistic imaging\, anti-count erfeiting). He directs the Visible Human Web Server project\, which offers advanced visualization services for exploring human anatomy.Fore more inf ormation\, get the DPS paper at http://diwww.epfl.ch/w3lsp/publications/gi gaserver/\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a033091 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033091 END:VEVENT BEGIN:VEVENT SUMMARY:Implementing Web Services DTSTART;VALUE=DATE-TIME:20030813T140000Z DTEND;VALUE=DATE-TIME:20030813T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033092@cern.ch DESCRIPTION:The topic of Web Services has been mentioned a great deal in t he computing press. As EDH is a fully-compliant J2EE application developed here at CERN\, and as part of an on-going activity to provide automated a ccess to EDH\, one of our developers has been working in collaboration wit h AB Division to set up an interface to part of EDH using Web Services. In this IT Seminar we will give a brief introduction to Web Services\, follo wed by our practical experiences gained from developing a Web Service in J ava. We will also describe how we overcame the interoperability issues whe n using the Web Service from an application written in Visual Basic. We wi ll end the seminar by outlining our future plans for a small e-commerce We b Service that we have in the pipeline.Organiser(s): Julian Blake / IT Div isionMore information: http://wwwinfo.cern.ch/seminars/\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=a033092 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033092 END:VEVENT BEGIN:VEVENT SUMMARY:New Applications of Digital Image Processing in Technology and Med icine DTSTART;VALUE=DATE-TIME:20030815T080000Z DTEND;VALUE=DATE-TIME:20030815T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033093@cern.ch DESCRIPTION:In the seminar an overview of image processing hardware develo pment and basic algorithms is given. The various image sensors used in tec hnology and medicine are explained. Examples of image processing applicati ons in automated assembly\, sound track restoration\, computer based learn ing and 3D navigation in medicine are presented.Organiser(s): Julian Blake / IT DivisionMore information: http://wwwinfo.cern.ch/seminars/\n\nhttp:/ /indico.cern.ch/conferenceDisplay.py?confId=a033093 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033093 END:VEVENT BEGIN:VEVENT SUMMARY:Database centric printer spool management and what this means to C ERN users DTSTART;VALUE=DATE-TIME:20030820T140000Z DTEND;VALUE=DATE-TIME:20030820T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a033094@cern.ch DESCRIPTION:Over the last year the CERN print service has undergone a majo r upgrade to the servers in terms of hardware and software. What these cha nges involve\, what the current state is\, and how they effect the CERN us ers\, are the subjects of this talk.\nWith around 15\,000 jobs per day on 1200 printers and a total of 2000 queues\, a scalable and robust spool man agement system is a necessity. With synchronisation and scalability proble ms in the past\, the new system has been designed around a central databas e that acts as an intermediary between the management interface and config uration generation programs.\nThe new management interface has many advant ages both to the support teams and to end users. Changes in the back-end f ilters and a new version of XPrint have brought new functionality to Unix and Linux users. The interface and tools will be presented in the talk wit h the hope that end users will be able to know about\, and fully utilise a ll of the available features.Organiser(s): Julian Blake / IT DivisionMore information: http://wwwinfo.cern.ch/seminars/\n\nhttp://indico.cern.ch/con ferenceDisplay.py?confId=a033094 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a033094 END:VEVENT BEGIN:VEVENT SUMMARY:The Future of the Grid and Globus: Beyond Web Services and OGSI DTSTART;VALUE=DATE-TIME:20030826T120000Z DTEND;VALUE=DATE-TIME:20030826T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a035144@cern.ch DESCRIPTION:Globus Toolkit(R) v3.0 was released in June 2003 with a comple te implementation of the Web services-based Open Grid Services Infrastruct ure (OGSI) v1.0 specification\, and a set of "Globus proprietary" Web/Grid services built on top of this OGSI basis. Our focus for the next 12-18 mo nths is to standardize the functionality available in Globus Toolkit\, as well as additional core\, complementary functionality\, and to make this a vailable in a future version of the Globus Toolkit. With this foundation i n place\, we then expect the emergence of a rich ecosystem of collective l ayer services that exploit this common foundation. In this talk I will ove rview the current GT3 structure\, explain the standards landscape to descr ibe how I think these standards and the ensuing GT implementation will evo lve\, and discuss the transition from GT2 into this evolving GT3 setting. Emerging standards to be discussed include WSDL (W3C)\, OGSI (GGF)\, OGSI- Agreement (GGF)\, Web Services Distributed Management and Web Services Man agement Framework (Oasis)\, Data Access and Integration (GGF)\, data repli cation and federation\, events and message queues\, and job submission and reservation.BiographySteve Tuecke (pronounced Tee-key) is lead architect of the Globus Project(tm)\, responsible for the architecture and design of the Globus Toolkit(R) and for the Grid and Web Services standards that un derlie it. He began working as a software developer for Ian Foster in 1990 in the Mathematics and Computer Science Division at Argonne National Labo ratory. Since the beginning of the Globus Project in 1996\, Tuecke has bee n responsible for building Foster's Distributed Systems Laboratory (DSL) a t Argonne virtually from scratch into the premier Grid research and develo pment group in the world\, including staffing\, management\, and technolog y leadership. Since turning over management of the DSL in 2001\, Tuecke ha s focused on Globus Toolkit architecture and design\, Grid standards\, and industry relationships.\nTuecke graduated Summa Cum Laude from St. Olaf C ollege in 1989\, receiving a B.A. in Mathematics and Computer Science. In 2002 he received Technology Review magazine's TR100 award\, recognizing hi m as one of the world's top 100 young innovators. In 2002 he also was name d to Crain's Chicago Business Forty Under 40\, as one of the Chicago area' s "best and brightest". In 2003 he was named (with Foster and Kesselman) b y InfoWorld Magazine as one of its Top 10 Technology Innovators of 2003\, as the "masterminds behind the technologies that are making the biggest sp lash and steering IT into the future". He is a leader in the Grid and Web services standards communities\, including a member of the Global Grid For um Steering Group\, editor and co-chair of the Open Grid Services Infrastr ucture (OGSI) standard\, and an author or participant of numerous other st andards efforts including W3C WSDL (Web Services)\, GGF OGSI-Agreement (SL A negotiation)\, GGF Data Access and Integration\, OASIS WSDM (service man agement)\, and IETF X.509 Proxy Certificates (security).Organiser(s): Bob Jones / IT DivisionInformation: http://wwwinfo.cern.ch/seminars/\n\nhttp:/ /indico.cern.ch/conferenceDisplay.py?confId=a035144 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a035144 END:VEVENT BEGIN:VEVENT SUMMARY:Performance test of PCs based on AMD platforms DTSTART;VALUE=DATE-TIME:20030903T140000Z DTEND;VALUE=DATE-TIME:20030903T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a035128@cern.ch DESCRIPTION:ALICE at CERN will be one of the most demanding physics experi ments in the world concerning data throughput\, data processing and data v olumes. In order to find the best choice of computing machinery to acquire and process these data\, many platforms are under evaluation at the momen t. This talk reports on the testing of a rack mountable system based on AM D 32 bit dual AthlonMP processors and on AMD 64 bit Opteron machines\, on loan from ICT AG\, thanks to the help of AMD Europe in Geneva. The benchma rk tests and the results obtained will be described.Organiser(s): Jamie Sh iers / IT DivisionMore information: http://wwwinfo.cern.ch/seminars/\n\nht tp://indico.cern.ch/conferenceDisplay.py?confId=a035128 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a035128 END:VEVENT BEGIN:VEVENT SUMMARY:Building an Open Grid: A Status Report DTSTART;VALUE=DATE-TIME:20030909T120000Z DTEND;VALUE=DATE-TIME:20030909T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a035216@cern.ch DESCRIPTION:Grid technologies and infrastructure support the integration o f services and resources within and among enterprises\, and thus allow new approaches to problem solving and interaction within distributed\, multi- organizational collaborations. Sustained effort by computer scientists and application developers has resulted in the creation of a substantial open source technology\, numerous infrastructure deployments\, a vibrant inter national community\, and significant application success stories. Long-ter m success now depends critically on three issues: open standards\, open so ftware\, and open infrastructure. I discuss current efforts and future dir ections in each area\, referring in particular to recent developments in " cyber-infrastructure" in the U.S.\, EGEE in Europe\, Open Grid Services Ar chitecture standards\, and adoption within scientific scientific communiti es beyond the physical sciences.BiographyDr. Ian Foster is Associate Direc tor of the Mathematics and Computer Science Division at Argonne National L aboratory and Professor of Computer Science at the University of Chicago. The Distributed Systems Lab that he heads at Argonne and Chicago is home t o the Globus Toolkit\, the open source software that has emerged as the de facto standard for Grid computing. He has published five books and over 2 00 articles and technical reports on various topics relating to programmin g languages\, parallel computing\, and distributed systems. He is a fellow of the British Computer Society\, and has received a number of awards for his research\, including the GII Next Generation Award and the British Co mputer Society's Lovelace Medal.See http://www-fp.mcs.anl.gov/~foster for further information.Organiser(s): Bob Jones / IT Division\n\nhttp://indic o.cern.ch/conferenceDisplay.py?confId=a035216 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a035216 END:VEVENT BEGIN:VEVENT SUMMARY:Oracle Developers Technical Session: Oracle XDB Based Reporting Fr amework DTSTART;VALUE=DATE-TIME:20031021T073000Z DTEND;VALUE=DATE-TIME:20031021T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a035734@cern.ch DESCRIPTION:XML has fast become a reference technology for exchanging and structuring data. Oracle offers a wide range of tools to generate\, manipu late\, validate and transform XML documents. With Oracle9i Database\, it i s now possible to develop efficient and easy to use XML based reporting fr ameworks. \nThe purpose of this session is to show you such a framework de veloped by Trivadis which extracts data from relational tables or views\, transforms the XML result set into various output types\, such as HTML\, P DF\, Excel\, SVG or XML using XML stylesheets stored in the Database.\nTh e framework generates web reports and maintenance forms (e.g. for updates\ , inserts and deletes).\nTo demonstrate the framework functionality a repo rt will be created in front of the audience using XMLspy.Organiser(s): Ped ro Martel / EST Division\n\nhttp://indico.cern.ch/conferenceDisplay.py?con fId=a035734 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a035734 END:VEVENT BEGIN:VEVENT SUMMARY:Oracle Developers Forum DTSTART;VALUE=DATE-TIME:20031201T090000Z DTEND;VALUE=DATE-TIME:20031201T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a036345@cern.ch DESCRIPTION:Agenda10:00Oracle Training NewsArash Khodabandeh (IT/DB)10:15 Report from Oracle Open World conferences (San Francisco & Paris)Nilo Segu ra et al. (IT/DB)10:45Coffee Break11:00Main features in 9iR2 + Migration p lansNilo Segura et al. (IT/DB)Organiser(s): Pedro Martel / EST Division\n\ nhttp://indico.cern.ch/conferenceDisplay.py?confId=a036345 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a036345 END:VEVENT BEGIN:VEVENT SUMMARY:AskOnce: the software to federate multiple search engines DTSTART;VALUE=DATE-TIME:20031202T143000Z DTEND;VALUE=DATE-TIME:20031202T153000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a036423@cern.ch DESCRIPTION:AskOnce is an Enterprise Content Integration solution that mak es global knowledge accessible with a single query.\nIt assimilates conten t from all corporate and deep web sources\, as well as enterprise applicat ions. Selected content can subsequently be integrated into other applicati ons such as databases\, spreadsheets\, groupware tools\, etc.\nHow does it work?\nSimply enter your search and askOnce goes to work\, automatically logging on to each internal or external data source\, rapidly sifting thro ugh massive amounts of data and returning the precise information you need .\nThe utility will be specifically explained in CERN context\, showing ho w various internal and external document servers could be searched for wit h a single query.Organiser(s): CERN Library & CDSMore information: http:// www.askonce.com\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a0364 23 LOCATION:CERN TH Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a036423 END:VEVENT BEGIN:VEVENT SUMMARY:Data Pump in Oracle Database 10g: Foundation for Ultra-High Speed Data Movement Utilities DTSTART;VALUE=DATE-TIME:20040205T093000Z DTEND;VALUE=DATE-TIME:20040205T103000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a04394@cern.ch DESCRIPTION:The Data Pump is a replacement for the Oracle export and impor t utilities. It is used to migrate portions of a database between database instances. It is able to achieve high performance through the management of parallel streams of execution. It also contains many desirable features that were missing in the old export and import utilities.\n\nAbout the sp eaker\n\nBill Fisher has been at Oracle for over 9 years developing databa se utilities. He implemented a parallel loader for RDB\, led the export/im port projects for several years\, and most recently has been one of the te am leaders for the Data Pump project. Prior to working at Oracle\, Bill wo rked at Digital Equipment Corporation for 19 years in the areas of operati ng system development\, networking\, semiconductor computer aided design\, typesetting\, and office automation. Bill lives in New Hampshire in the United States where he spends a lot of time avoiding calls from presidenti al election pollsters.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confI d=a04394 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a04394 END:VEVENT BEGIN:VEVENT SUMMARY:The Swiss Centre for Scientific Computing (CSCS): Accelerating Sci entific Discovery DTSTART;VALUE=DATE-TIME:20040325T150000Z DTEND;VALUE=DATE-TIME:20040325T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a041076@cern.ch DESCRIPTION:The CSCS mission has been recently reinforced by the adoption of a development and investment plan\, more autonomy and a new governance: its mission is to provide the calculating power\, storage capacity and ex pertise for the simulation\, modelling and analysis of complex data\, nece ssary for Switzerland's leading scientific research. In 2003\, in collabor ation with CERN and the Swiss particle physics research laboratories\, the centre began the construction of the Swiss link (Tier 2) of the worldwide LCG Grid. In the coming weeks CSCS\, together with its partners\, will be launching the national Grid Initiative for a selection of applications be yond particle physics.\n\nThis talk will give the perspective of the devel opment plan of the CSCS for the 4 years to come. Localization\, integrity and access to data\, computing resources and skills have become steadily m ore important in scientific computing. In this context\, the CSCS has a ne w role to play as facilitator for interdisciplinary collaborations of nati onal scope in areas such as deep computing\, distributed computing\, bench marking\, data-intensive computing and remote visualization.\n\nThe portfo lio of scientific applications at CSCS is presently formed by molecular sc iences\, computational chemistry\, material modelling\, climate and meteor ology (partner of Meteo Swiss for daily forecast)\, engineering\, physics and high energy physics\, and is rapidly opening up to new disciplines\, l ike biomedical applications.\n(http://www.cscs.ch/)\n\nOrganiser(s): Julia n Blake / IT Department\nMore information: http://cern.ch/computing-semina rs\n\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp://i ndico.cern.ch/conferenceDisplay.py?confId=a041076 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a041076 END:VEVENT BEGIN:VEVENT SUMMARY:InDiCo Workshop DTSTART;VALUE=DATE-TIME:20040423T120000Z DTEND;VALUE=DATE-TIME:20040423T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a041551@cern.ch DESCRIPTION:The INtegrated DIgital COnferencing EU project has finished bu ilding a complete software solution to facilitate the MANAGEMENT OF CONFER ENCES\, workshops\, schools or simple meetings from their announcement to their archival.\n\n\nSEE THE WORKSHOP AGENDA !\n\n\nInDiCo Headlines:\ncre ate and maintain your conference web site\, be it a single talk or a whole workshop\nschedule sessions and contributions\nattach material to confere nces and contributions\, be it text or multimedia content\nmake categories to organise conferences according to a hierarchy\nreview papers on the we b and email notifications to track coordinators\nprotect items based on us er authentication\ndelegate responsibilities to others according to roles\ narchive your conference to specialised repository services\n\n==> InDiCo is OAI-Compliant\n==> InDiCo will become the next version of CERN CDS Agen da (http://agenda.cern.ch)\n==> InDiCo is already used by CHEP 2004 confer enceEverybody involved in the organization of events is welcome to join th is workshop. More information about InDiCo can be found at http://indico.s issa.it and http://cern.ch/indico\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=a041551 LOCATION:CERN 60-6-002 URL:http://indico.cern.ch/conferenceDisplay.py?confId=a041551 END:VEVENT BEGIN:VEVENT SUMMARY:The OurGrid Project: Grid Computing for Bag-of-Tasks Applications DTSTART;VALUE=DATE-TIME:20040504T120000Z DTEND;VALUE=DATE-TIME:20040504T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a041661@cern.ch DESCRIPTION:Grid Computing has appeared with the enticing promise of turni ng computing into utility. The vision is 'plug in the grid and solve your problem'. However\, turning the grid vision into reality is no trivial mat ter. Despite the great progress made in the last years\, Grid Computing is still far from reality to most users. The OurGrid Project aims to deliver grid technology that can be used today by current users to solve present problems. To achieve this goal\, OurGrid chooses a different trade-off com pared to most grid projects. It forfeits supporting arbitrary applications in favor of supporting only Bag-of-Tasks applications. Bag-of-Tasks appli cations are those parallel applications whose tasks are independent of eac h other. Despite their simplicity\, BoT applications are used in a variety of scenarios\, including data mining\, massive searches\, parameter sweep s\, monte-carlo simulations\, fractal calculations\, computational biology \, and computer imaging. The OurGrid solution has three major components: the MyGrid broker\, the OurGrid community\, and SWAN sandboxing. MyGrid is a broker for Bag-of-Tasks applications that allows a user to run her appl ication on all processors she has access to. The OurGrid community is a pe er-to-peer network designed to enable worldwide resource sharing\, without relying upon grid economy infrastructure (which is not deployed today). S WAN leverages the Bag-of-Tasks characteristic of the application to provid e sandboxing that allows sites to safely run guest. grid applications. All OurGrid components are open source. MyGrid has been available for 1.5 yea rs and has a user community of around 50 users. OurGrid is in prototype st age\, deployed in a small-scale experimental community. SWAN is in late de velopment stage.\n\nOrganiser(s): Julian Blake / IT Department\nMore infor mation: http://cern.ch/computing-seminars\n\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a041661 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a041661 END:VEVENT BEGIN:VEVENT SUMMARY:Stork: Making Data Placement a First Class Citizen in the Grid DTSTART;VALUE=DATE-TIME:20040525T090000Z DTEND;VALUE=DATE-TIME:20040525T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a042110@cern.ch DESCRIPTION:Data placement is an essential part of today's Grid applicatio ns. Moving the data close to the application for efficiency\, and replicat ing the data for reliability\, are crucial. The increasing data requiremen ts of both scientific and commercial applications\, and collaborative acce ss to these data make this problem even more important. In the current app roach\, data placement is regarded as a side effect of computation. Our go al is to make data placement a first class citizen in the Grid just like t he computational jobs. They will be queued\, scheduled\, monitored\, manag ed\, and even check-pointed. Since data placement jobs have different char acteristics than computational jobs\, they cannot be treated in the exact same way as computational jobs are treated. For this purpose\, we are prop osing a framework that can be considered as a "data placement subsystem" for the Grid\, similar to the I/O subsystem in operating systems. This fra mework includes a specialized scheduler for data placement\, a high level planner aware of data placement jobs\, a resource broker/policy enforcer a nd some optimization tools. The proposed system can perform reliable and e fficient data placement\, it can recover from all kinds of failures withou t any human intervention\, and it can dynamically adapt to the environment at the execution time.\n\nOrganiser(s): Julian Blake / IT Department\nMor e information: http://cern.ch/computing-seminars\n\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisp lay.py?confId=a042110 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a042110 END:VEVENT BEGIN:VEVENT SUMMARY:Virtual Reality and Visualization: recent progress in acquiring an d displaying 3D scenes and multiple data sets DTSTART;VALUE=DATE-TIME:20040601T120000Z DTEND;VALUE=DATE-TIME:20040601T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a042107@cern.ch DESCRIPTION:Virtual reality (VR) is becoming a much more affordable and ro utine technology\, increasing its availability and usefulness for science. This talk will describe three challenges where VR-related solutions could provide benefits to large-scale experiments such as the LHC.\n\nThe first challenge concerns providing a virtual model of a large-scale\, inaccessi ble facility\, for example for the general public to experience the facili ty remotely. This challenge involves effectively acquiring a large 3D mode l of the facility. Two methods of acquiring 3D models of large\, complicat ed interior spaces will be discussed: a commercially-available solution us ing a laser range finder and calibrated camera and a research system using an uncalibrated freely-moving video camera.\n\nThe second challenge conce rns how to create a convincing virtual environment for a very large facili ty. User studies showing the importance of 'static haptics' and locomotion by walking in virtual environments will be presented\, along with a metho d of 'redirected walking' that enables exploring virtual spaces much large r than the physical space used.\n\nThe third challenge concerns how to com prehend output from a complex experiment\, which requires effective displa y of many variables. The best ways to display multiple scalar and vector f ields on surfaces will be presented\, and the promising but ultimately unf ruitful methods that have been tested in the course of finding optimum sol utions will be discussed.\n\nAbout the speaker:\nRussell Taylor is a Resea rch Associate Professor of Computer Science\, Physics & Astronomy\, and Ap plied & Materials Sciences at the University of North Carolina at Chapel H ill. He is the co-director of the http://www.cs.unc.edu/Research/nano/cism m>UNC NIH National Research Resource for Computer Integrated Systems for M icroscopy and Manipulation (come visit!).\n\nHis research interests includ e Scientific Visualization\, Distributed Virtual Worlds\, Haptic Display\, and Interactive 3D Computer Graphics. All of these come together in his r ole as the director of the computer science team in the UNC http://www.cs. unc.edu/Research/nano>Nanoscale Science Research Group\, which is a team o f Physicists\, Chemists\, Gene Therapists\, Biologists\, Library Scientist s\, Perceptual Psychologists\, and Computer Scientists working together to develop improved interfaces for scanned-probe and other microscopes. See www.cs.unc.edu/~taylorr.\n\nhttp://indico.cern.ch/conferenceDisplay.py?con fId=a042107 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a042107 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Edinburgh HEPiX-HEPNT Meeting DTSTART;VALUE=DATE-TIME:20040604T073000Z DTEND;VALUE=DATE-TIME:20040604T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a042029@cern.ch DESCRIPTION:The\nSpring 2004 meeting of HEPiX-HEPNT\nwill be held in Edinb urgh from 24th to 28th May 2004. HEPiX is a global organization of service managers and support staff providing computing facilities for the High En ergy Physics community. All operating systems used by HEP are covered\, in cluding UNIX\, Windows and Grid computing. The first three days of the mee ting will be devoted to general HEP computing issues. The last one and a h alf days are for a Mass Storage workshop.\n\nThe speakers at this joint Co mputing/After-C5 Seminar will present their views of the HEPiX meeting.\n\ nOrganiser(s): Julian Blake / IT Department\nMore information: http://cern .ch/computing-seminars\n\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Depa rtment\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a042029 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a042029 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Interlaken CHEP'04 Conference DTSTART;VALUE=DATE-TIME:20041014T140000Z DTEND;VALUE=DATE-TIME:20041014T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a044233@cern.ch DESCRIPTION:This year's\nComputing in High Energy and Nuclear Physics conf erence is being held in Interlaken from 27th September to 1st October 2004 . CHEP conferences provide an international forum to exchange information on computing experience and needs for the High Energy Physics and Nuclear Physics communities\, and to review recent\, ongoing and future activities .\n\nThe speakers at this Computing Seminar will present their views of th e CHEP'04 conference.\n\nOrganiser(s): Julian Blake / IT Department\nMore information: http://cern.ch/computing-seminars\n\n© CERN 2005 -\nMiguel A ngel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=a044233 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a044233 END:VEVENT BEGIN:VEVENT SUMMARY:BSD UNIX in Mac OS X DTSTART;VALUE=DATE-TIME:20041103T150000Z DTEND;VALUE=DATE-TIME:20041103T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a044579@cern.ch DESCRIPTION:The most widely distributed UNIX-based operating system\, Mac OS X offers a unique combination of technical elements to the discerning g eek\, such as the fine-grained multithreading of the Mach 3.0 kernel\, tig ht hardware integration and SMP-safe drivers\, as well as zero configurati on networking.\n \nMac OS X integrates features from state-of-the-art Fre eBSD 5 into Darwin - the Open Source\, UNIX-based foundation of Mac OS X - to provide enhanced performance\, compatibility and usability.\n\nJordan Hubbard will present:\n its UNIX lineage and primary reference code bases\ ,\n challenges of marrying UNIX with an ease-of-use doctrine\,\n Enterpris e versus Desktop challenges\,\n what Apple has done so far in the UNIX are na\,\n what Apple plans to do in the future and why.\n About the speaker:\ nJordan Hubbard\, the manager of BSD Technologies for Apple\, keeps a tigh t lid over the chaos that could be Darwin (the UNIX-based core of the Mac OS X). Before arriving at Apple\, he was the Principal Technologist for Wi nd River Systems and was responsible for the FreeBSD CD-ROM product line. Jordan co-founded and ran the FreeBSD project and still contributes to the Open Source community. He's held various engineering and management posit ions in the U.S. and Europe\, having begun his career working as a teen on minicomputers back in the 1970's.\n\nOrganiser(s): Julian Blake / IT Depa rtment\nNote: http://cern.ch/computing-seminars\n© CERN 2005 -\nMiguel An gel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay .py?confId=a044579 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a044579 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Brookhaven HEPiX Meeting DTSTART;VALUE=DATE-TIME:20041105T083000Z DTEND;VALUE=DATE-TIME:20041105T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a044698@cern.ch DESCRIPTION:The \nAutumn 2004 meeting of HEPiX-HEPNT\nwas held at Brookhav en from 18th to 22nd October 2004. HEPiX is a global organization of servi ce managers and support staff providing computing facilities for the High Energy Physics community. All operating systems used by HEP are covered\, including UNIX\, Windows and other operating environments. The first three days of the meeting were devoted to general HEP computing issues. The las t one and a half days were for a workshop with topic 'Large Systems SIG / Platforms for Physics'.\n\nThe speakers at this joint Computing/After-C5 S eminar will present their views of the HEPiX meeting.\n\nOrganiser(s): Jul ian Blake / IT Department\nMore information: http://cern.ch/computing-semi nars\n\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp:/ /indico.cern.ch/conferenceDisplay.py?confId=a044698 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a044698 END:VEVENT BEGIN:VEVENT SUMMARY:ITIL\, Best Practice\, and IT Service Management DTSTART;VALUE=DATE-TIME:20041109T130000Z DTEND;VALUE=DATE-TIME:20041109T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a044699@cern.ch DESCRIPTION:The IT Infrastructure Library (ITIL) is a framework for best p ractices for IT Service Management. ITIL was developed and matured in Euro pe and is today an international de facto standard applicable to a very wi de range of organizations. The objective of this seminar is to describe th e scope of IT Service Management\; to investigate its goals and its benefi ts\; and to describe how CERN may take advantage of the best practice appr oach. Particular attention will be paid to the challenges existing at CERN today\, as well as the particular benefits to CERN in the future\, when s uch services as Grid computing enter the production phase.\n\nSpeakers:\nM r Seymour Hosking\, Hosking Associates\, and Mr Robert Falkowitz\,\nConcen tric Circle Consulting.\n\nOrganiser(s): Julian Blake / IT Department\nMor e information: http://cern.ch/computing-seminars\n© CERN 2005 -\nMiguel A ngel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=a044699 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a044699 END:VEVENT BEGIN:VEVENT SUMMARY:Practical Use of XML DTSTART;VALUE=DATE-TIME:20041201T150000Z DTEND;VALUE=DATE-TIME:20041201T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a045167@cern.ch DESCRIPTION:Extensible Markup Language (XML) is becoming more and more pop ular for data exchange\, storage and presentation. It is also proven to b e very useful for defining other data handling formats\, such as WSDL or S VG.\n \nThe speaker will give an overview of XML and its related technolo gies - XML-schemas\, XPath\, XSLT\, XSL:FO - and provide examples of its p ractical use at CERN in the context of the Electronic Document Handling sy stem (EDH) project.\n\nAbout the speaker:\nRostislav ("Slava") Titov first arrived at CERN in 1995 and since then he works as a software engineer in various administrative information systems projects at CERN. For the las t 5 years Slava was one of the core developers of the CERN Electronic Docu ment Handling system (EDH).\n\nOrganiser(s): Julian Blake / IT DepartmentM ore information: http://cern.ch/computing-seminars\n\n© CERN 2005 -\nMigu el Angel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDi splay.py?confId=a045167 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a045167 END:VEVENT BEGIN:VEVENT SUMMARY:Computing with Floating Point: - It's not Dark Magic\, it's Scienc e DTSTART;VALUE=DATE-TIME:20050111T130000Z DTEND;VALUE=DATE-TIME:20050111T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a05121@cern.ch DESCRIPTION:Too many programmers and scientists would define floating-poin t numbers as 'some kind of fuzzy approximation to the reals in computers\, which is usually accurate enough - and when it's not\, bad luck'. Unfortu nately\, this was a fairly accurate description until 1985.\n\nThe situati on then changed with the IEEE-754 standard for floating-point arithmetic. This standard defines three floating-point formats\, four rounding modes\, and the complete behaviour of the operators +\, -\, x\, / and square root . In principle\, a program involving only these operations is deterministi c\, portable to any IEEE-754 system\, and it becomes possible to prove pro perties of its numerical behaviour: In short\, Science shall prevail. And it is fortunate\, as floating-point is increasingly used in airliner fligh t computers and\, it seems\, high-energy scientific equipment.\n\nThe purp ose of this seminar is therefore to convince real-world scientists to appr oach floating-point with a scientific view. We will first show you that po rtability is good\, but does not guarantee accuracy: in some cases\, float ing-point will eat all your precision without warning. We will show you ho w to avoid\, or at least detect these traps. We will give examples of floa ting-point based numerical proofs\, and if this frightens you\, there will be more experimental approaches to improving the confidence in your float ing-point code.\n\nIn 2004\, however\, IEEE-754 compliance still required a tight and harmonious cooperation between a language\, a compiler\, an op erating system and a processor\, which is far from being automatic. We wil l survey these questions. Among the responsibilities of the operating syst em\, the elementary functions (exp\, log\, the trigonometric functions\, e tc) were left out of the standard. Our current research goal is to bring t hem in\, and we will explain why it is difficult\, and how it is possible. \n\nAbout the speaker:\nFlorent de Dinechin received his DEA from the Éco le Normale Supérieure de Lyon (ENS-Lyon) in 1993\, and his PhD from Unive rsité de Rennes 1 in 1997. After a postdoctoral position at Imperial Coll ege\, London\, he is now a permanent lecturer at ENS-Lyon in the Arénaire team of the Laboratoire de l'Informatique du Parallélisme (LIP). His res earch interests include computer arithmetic\, software and hardware evalua tion of functions\, computer architecture and FPGAs.\n\n\nOrganiser(s): \n Julian Blake\nComputing Seminars /\nIT Department\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=a05121 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a05121 END:VEVENT BEGIN:VEVENT SUMMARY:Web Services for Grid Computing - WSRF and WSRF::Lite DTSTART;VALUE=DATE-TIME:20050117T150000Z DTEND;VALUE=DATE-TIME:20050117T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a05122@cern.ch DESCRIPTION:WSRF::Lite is an implementation of the Web Service Resource Fr amework (WSRF\, http://www.globus.org/wsrf/) in Perl. It supersedes OGSI:: Lite\, the Perl implementation of the Open Grid Service Infrastructure (OG SI\, https://forge.gridforum.org/projects/ogsi-wg). Perl is a powerful too l for developing Web services and Web service clients. WSRF::Lite illustra tes this by providing a relatively simple environment for developing WS-Re sources. The seminar will discuss WSRF\, building Web services in Perl and using WSRF::Lite.\n\nAbout the speaker:\nMark obtained his PhD in Atomic Physics from Queen's University of Belfast. He first became acquainted wit h grid computing through the Legion Project while working as Research Supp ort Consultant at the University of Virginia in 1999. He later joined the University of Manchester\, where he developed OGSI::Lite while working for the UK Grid Support Centre. He released the world's first publicly availa ble implementation of WSRF in February 2004\, within weeks of the announce ment of WSRF. He is currently continuing the development of WSRF::Lite thr ough the Managed Programme of the Open Middleware Infrastructure Institute (http://www.omii.ac.uk).\n\n\nOrganiser(s): \nJulian Blake\nComputing Sem inars /\nIT Department\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Depart ment\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a05122 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a05122 END:VEVENT BEGIN:VEVENT SUMMARY:Square Kilometre Array international radio telescope DTSTART;VALUE=DATE-TIME:20050121T090000Z DTEND;VALUE=DATE-TIME:20050121T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a05328@cern.ch DESCRIPTION:The\nSquare Kilometre Array (SKA)\nis a global project to desi gn and build a new generation radio telescope at metre to centimetre wavel engths. It will have a collecting area of order one million square metres spread over 3000 km\, a sensitivity 100 times higher than the current mos t sensitive interferometer\, an instantaneous field of view (FOV) of at le ast 1 square degree and\, in some designs\, more than one FOV allowing mul tiple simultaneous use. It will be an extremely powerful survey telescope with the capability to follow up individual objects with high angular and time resolution. The SKA science impact will be felt in astro-particle p hysics and cosmology\, fundamental physics\, galactic and extragalactic as tronomy\, and solar system science. Technological innovation\, closely pa ralleling commercial IT developments\, is key to the design concepts under investigation and to the cost goal of 1000 Euro/m2. Data transport rates are likely to be in the range of tera-bits/sec\, with Pflops capacity req uired for the central processor. The selection of technologies for the SK A is scheduled in early 2009. A number of possible locations for the tele scope are under investigation with a choice scheduled in late 2006. Const ruction of the array will take most of the next decade.\n\nAbout the speak er:\nRichard Schilizzi has been International Project Director for the Squ are Kilometre Array Project since 1 January 2003. For the ten years previo us to this he was foundation Director of the Joint Institute for Very Long Baseline Interferometry in Europe. He is also a professor of astronomy a t the University of Leiden in the Netherlands.\n\n\nOrganiser(s): \nJulian Blake\nComputing Seminars /\nIT Department\n© CERN 2005 -\nMiguel Angel Marquina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a05328 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a05328 END:VEVENT BEGIN:VEVENT SUMMARY:Comparison of IPv4 and IPv6 DTSTART;VALUE=DATE-TIME:20050222T150000Z DTEND;VALUE=DATE-TIME:20050222T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a051082@cern.ch DESCRIPTION:IP version 6 is a new IP protocol designed to replace IP versi on 4\, the Internet protocol that is predominantly deployed and extensivel y used throughout the world. The current version of IP has not been substa ntially changed since RFC 791\, Internet Protocol DARPA Internet Program P rotocol Specification\, was published in 1981. IPv4 has proven to be robus t\, easily implemented\, and interoperable\, and has stood the test of sca ling an internetwork to a global utility the size of the Internet today. H owever\, the initial design did not anticipate the following conditions:\n \nRecent exponential growth of the Internet and the impending exhaustion o f the IPv4 address space.\nGrowth of the Internet and the ability of Inter net backbone routers to maintain large routing tables.\nNeed for simpler a utoconfiguration and renumbering.\nRequirement for security at the IP leve l.Need for better support for real-time delivery of data\, also called qua lity of service (QoS).\nNote: Features such as IP Security (IPSec) and QoS have been specified for both versions of IP. The session will discus the changes and mechanics between IPv4 versus IPv6 to address the ongoing prog ress of the Internet.\n\nAbout the speaker:\nGunter Van de Velde is a Seni or Network Consulting Engineer in Cisco's Advanced Services Team. He has b een the Technical Team Leader for the IP Core delivery team since early 20 01. Gunter received his Master degree in Electronics in 1993. His first pr ofessional activities where based on TDMs\, modems and L2-bridges. He join ed Cisco in 1997\, initially providing reactive worldwide support as part of Cisco's Technical Assistence Center (also known as 'Cisco TAC')\, speci alizing in IP Routing protocol technologies. In 1999 he joined Cisco's Adv anced Services organization as a Network Consultant\, where he has been ac tive in designing large Backbone ISP networks and Services. Since 2001 Gun ter has been working as Design Architect for the 6NET IPv6 project. Since then Gunter has been involved in optimizing and supporting the 6NET Infras tructure - his latest project on 6NET is to optimize the overall routing c onvergence on the 6NET European backbone.\n\n\nOrganiser(s): \nJulian Blak e\nComputing Seminars /\nIT Department\n© CERN 2005 -\nMiguel Angel Marqu ina /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confI d=a051082 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a051082 END:VEVENT BEGIN:VEVENT SUMMARY:A peek into the world of chip design DTSTART;VALUE=DATE-TIME:20050315T150000Z DTEND;VALUE=DATE-TIME:20050315T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a051490@cern.ch DESCRIPTION:This lecture will give some insight into how the Microprocesso r Design Group approaches the daunting task of the design of a lead microp rocessor as complex as the Pentium IV while under very specific schedule c onstraints. For a historical perspective\, we will start with a quick comp arison of the complexity/performance of the Willamette and Prescott (Penti um IV) class of microprocessor with the generations before the Pentium IV (Bob was a member of the design teams for a number of lead microprocessor projects including the 486DX2\, Pentium Pro\, Willamette Pentium IV and th e Prescott Pentium IV). This historical perspective discussion will set th e stage for a discussion of some of the details and methodology of the pro ject culture and design engineering culture that were critical to the succ ess of the design efforts for the Willamette and Prescott projects. The Ve rtical Design Engineer Task Flow Diagram will then be discussed in detail. The insight presented in this lecture should be of benefit and of interes t to anyone involved in a large technical project.\nAbout the speaker\nPri or to his retirement from Intel in Oct 2002\, Bob spent 38 years in the fi eld of Electronic Design Engineering. For the last 18 years Bob was a Seni or CMOS Integrated Circuit Design Engineer with Intel's Microprocessor Des ign Group in Hillsboro\, Oregon USA. He was a major contributor involved i n the design of a number of lead microprocessor chips including the 486DX2 \, Pentium Pro\, Willamette Pentium IV and most recently the Prescott Pent ium IV (which is Intel's newest generation 32-bit microprocessor). Bob has extensive experience in nearly all aspects of microprocessor design inclu ding high-level language interpretation\, logic design\, circuit design\, memory design\, layout and prototype debug. He also has extensive experien ce in digital system/circuit design and analog circuit design.Prior to Int el\, from 1978 to 1984\, Bob was employed with Tektronix in Beaverton\, Or egon and was involved in CMOS and Ga-As integrated circuit design. From 19 69 to 1978 he was employed with Motorola Semiconductor Products Division i n Phoenix and Austin as a member of the Motorola CMOS Standard Products De sign Group. From 1965 to 1969 Bob was with the Motorola Government Electro nics Division in Scottsdale\, Arizona working on board level design of Apo llo and Mariner Mars electronic support systems. Over the years he has don e a significant amount of technical management and mentoring of Design Eng ineers. Bob holds 12 patents.\n\n\nOrganiser(s): \nMiguel Angel Marquina\n Computing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceD isplay.py?confId=a051490 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a051490 END:VEVENT BEGIN:VEVENT SUMMARY:Security in Computer Applications DTSTART;VALUE=DATE-TIME:20050322T150000Z DTEND;VALUE=DATE-TIME:20050322T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a051621@cern.ch DESCRIPTION:Computer security has been an increasing concern for IT profes sionals for a number of years\, yet despite all the efforts\, computer sys tems and networks remain highly vulnerable to attacks of different kinds. Design flaws and security bugs in the underlying software are among the ma in reasons for this.\n\nThis lecture addresses the following question: how to create secure software? The lecture starts with a definition of comput er security and an explanation of why it is so difficult to achieve. It th en introduces the main security principles (like least-privilege\, or defe nse-in-depth) and discusses security in different phases of the software d evelopment cycle. The emphasis is put on the implementation part: most com mon pitfalls and security bugs are listed\, followed by advice on best pra ctice for security development. The last part of the lecture covers some m iscellaneous issues like the use of cryptography\, rules for networking ap plications\, and social engineering threats.\n\nThis lecture was first giv en on Thursday\, February 24th\, 2005 at the Inverted CERN School of Compu ting 2005. Bibliography and further reading are available at\nhttp://cern. ch/Sebastian.Lopienski/Security\n\n\n\nOrganiser(s): \nMiguel Angel Marqui na\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/confere nceDisplay.py?confId=a051621 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a051621 END:VEVENT BEGIN:VEVENT SUMMARY:Data Logistics in Network Computing DTSTART;VALUE=DATE-TIME:20050418T140000Z DTEND;VALUE=DATE-TIME:20050418T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a052291@cern.ch DESCRIPTION:In distributed computing environments\, performance is often d ominated by the time that it takes to move data over a network. In the ca se of data-centric applications\, or Data Grids\, this problem of data mov ement becomes one of the overriding concerns. This talk describes techniqu es for improving data movement in Grid environments that we refer to as 'l ogistics.' We demonstrate that by using storage and cooperative forwardin g 'in' the network\, we can improve end to end throughput in many cases. Our approach offers clear performance benefits for high-bandwidth\, high-l atency networks. This talk will introduce the Logistical Session Layer (L SL) and provide experimental results from that system.\n\nAbout the speake r:\nMartin Swany is an Assistant Professor in the Department of Computer a nd Information Sciences at the University of Delaware. He completed his Ph .D. at the University of California\, Santa Barbara in 2003. He is a 2004 recipient of the US Department of Energy Early Career Principal Investigat or award. His research interests include distributed computing and network ing.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a052291 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a052291 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Karlsruhe HEPiX Meeting DTSTART;VALUE=DATE-TIME:20050527T073000Z DTEND;VALUE=DATE-TIME:20050527T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a053179@cern.ch DESCRIPTION:\n* Note unusual time!\n\n\nThe \nspring 2005 meeting of HEPiX \nwas held at Forschungszentrum Karlsruhe\n(Germany) from 9th to 13th May 2005. HEPiX is a global organization of\nservice managers and support staf f providing computing facilities for the\nHigh Energy and Nuclear Physics community. The first three days of the\nmeeting were devoted to general HE P computing issues. The last one and a\nhalf days were devoted to a topica l workshop on batch schedulers locally at\nsites and for the Grid.\n\nThe speakers at this joint Computing/After-C5 Seminar will present their\nview s of the HEPiX meeting.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComput ing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay .py?confId=a053179 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a053179 END:VEVENT BEGIN:VEVENT SUMMARY:Scalable XML programming techniques DTSTART;VALUE=DATE-TIME:20050615T140000Z DTEND;VALUE=DATE-TIME:20050615T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a053700@cern.ch DESCRIPTION:XML has silently taken over the world\, at least in the domain s of online/print publishing and applications that require exchanges of sm all amounts of data. However\, writing programs that do something with XML is still tedious\, because there are too many approaches that work only f or limited range of tasks. Examples are the use XSLT for recursive transfo rmations\, XQuery for database-like querying\, and conventional general pu rpose programs like Java that manipulate XML through clumsy library functi ons.\n\nScala is a general-purpose language that offers some remedy. The l anguage unifies functional and object-oriented programming\, which allows to easily integrate queries and transformations. In order to better suppor t XML programming\, the use of standard XML syntax inside Scala source cod e is permitted\, and a well-balanced library offers functionality\nthat is frequently needed for XML manipulation such as XPath selection. An experi mental subproject is data-binding\, which maps type information given in D TDs or an XML Schema into Scala classes.\n\nIn order to evaluate the usefu lness of the language\, we walk through the code of a real-life applicatio n that takes advantage of Scala's design: "aladdin" (http://maglite.epfl.c h/bugtracking ) is an interactive server application used at LAMP which co nnects user to a relational database\, generating XHTML on the fly. Apart from this application\, we have tools that translate XQuery and XSLT into Scala code in order to be able to migrate from these special-purpose lang uages. We discuss strengths and room for improvement.\n\nAbout the author: \nBurak Emir studied computer science in Aachen\, Germany (started 1997) a nd came to CERN as a summer student in 2000. He graduated in 2003\, and si nce then pursues a PhD degree at the Programming Languages and Methods Gro up at EPFL. His research topic are programming language abstractions for X ML data.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\ nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a053 700 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a053700 END:VEVENT BEGIN:VEVENT SUMMARY:Network computing with Einstein@home and climateprediction.net DTSTART;VALUE=DATE-TIME:20050711T140000Z DTEND;VALUE=DATE-TIME:20050711T153000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a054402@cern.ch DESCRIPTION:The\nBerkeley Open Infrastructure for Network Computing (BOINC )\nis an open-source platform that enables researchers to run distributed computing projects on thousands of public PCs. BOINC was developed by the team behind\nSETI@home\, the screensaver programme that analyses radioastr onomical data for signs of extraterrestrial intelligence\, which has been downloaded on more than 5 million computers. Since BOINC's launch two year s ago\, several scientific projects have implemented it\, attracting thous ands of 'PC volunteers'\, CERN's LHC@home project being one of these. This seminar will review two of these projects\, both from the point of view o f the scientific computing challenges they involve\, and the scientific re sults they aim to produce. The seminar is in two parts:\nEinstein@Home\nis a project developed to search data from the Laser Interferometer Gravitat ional wave Observatory (LIGO) in the US and from the GEO 600 gravitational wave observatory in Germany for signals coming from extremely dense\, rap idly rotating stars. Such sources are believed to be either quark stars or neutron stars\, and a subclass of these are already observed by conventio nal means as pulsars or X-ray emitting celestial objects. Einstein@home is an official project of the Year of Physics.\n\nclimateprediction.net\naim s to investigate the approximations that have to be made in state-of-the-a rt climate models by running such models thousands of times under a range of approximations. This will help understand how sensitive different model s are to small changes in\, for example\, the carbon dioxide and the sulph ur cycle. This will contribute to exploring how climate may change in the next century under a wide range of different scenarios.\nSpeakers:\nProf. Bruce Allen\,\nLIGO Scientific Collaboration Group\, University of Wiscons in-Milwaukee\nDr. Carl Christensen\, Computing Laboratory\, University of Oxford\nDr. Neil Massey\,\nComputing Laboratory\, University of Oxford\nDr . Tolu Aina\,\nAtmospheric\, Oceanic & Planetary Physics\, University of O xford\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT D epartment\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a054402 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a054402 END:VEVENT BEGIN:VEVENT SUMMARY:Java web hosting at CERN DTSTART;VALUE=DATE-TIME:20051101T150000Z DTEND;VALUE=DATE-TIME:20051101T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a056601@cern.ch DESCRIPTION:\n\nUp to September 2005\, there was no easy way to have a jav a web application hosted at CERN without having to run a J2EE server yours elf. The situation changed with the recent introduction of a new IT servic e: J2EE Public Service.\n\nWe first briefly introduce java web publishing technologies such as JSP and servlets. Then we show how these technologies can be used to prepare a java web application\, and how such application can be deployed and hosted within the service.\n\nFinally\, we discuss som e more advanced issues such as the implementation of\naccess restrictions using NICE authentication\, jdbc connectivity to oracle\ndatabases\, conne ction pooling and SOAP web services deployment.\n\n\nOrganiser(s): \nMigue l Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=a056601 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a056601 END:VEVENT BEGIN:VEVENT SUMMARY:LHC Commissioning Plans DTSTART;VALUE=DATE-TIME:20051110T100000Z DTEND;VALUE=DATE-TIME:20051110T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a056884@cern.ch DESCRIPTION:\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\ nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a056 884 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a056884 END:VEVENT BEGIN:VEVENT SUMMARY:BOINC: Architecture and Basic Principles DTSTART;VALUE=DATE-TIME:20051208T133000Z DTEND;VALUE=DATE-TIME:20051208T143000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a057901@cern.ch DESCRIPTION:BOINC is an open-source middleware to operate computing projec ts based on the Public Resource Computing paradigm. The best known is SETI @home\, the project from which BOINC is its spin-off. We are using it at C ERN to operate ours\, LHC@home (http://cern.ch/athome).\n\nLast month we h eld a workshop on the BOINC technology at CIEMAT\, Madrid in the framework of a collaboration between CERN and Extremadura. We present the material we prepared for the occasion\, explaining the basics of the architecture a nd the projects being under study for other applications of interest for L HC physics.\n\nRelated information\nThere have been two Computing Seminars on BOINC-related subjects:\nComputing with Floating Point: It's not Dark Magic\, it's Science - January 11\, 2005\nNetwork computing with Einstein@ home and climateprediction.net - July 11\, 2005\nAbout the speaker:\n\nJua n Antonio Lopez Perez obtained his Physics degree in the Valencia\nUnivers ity and has been working in IT at CERN from 2004. Among other\nprojects\, he has been involved in the porting of applications to the\nBOINC environm ent\, in the set up of a BOINC server and in the taught of\na BOINC semina r in Madrid.\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a0 57901 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a057901 END:VEVENT BEGIN:VEVENT SUMMARY:Enabling Data-Intensive Science with Tactical Storage Systems DTSTART;VALUE=DATE-TIME:20060125T150000Z DTEND;VALUE=DATE-TIME:20060125T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a057237@cern.ch DESCRIPTION:Large scale scientific computing requires the ability to share and consume data and storage in complex ways across multiple systems. How ever\, conventional systems constrain users to the fixed abstractions sele cted by the local system administrator.\n\nThe result is that users must e ither move data manually over the wide area or simply be satisfied with th e resources of a single cluster. To remedy this situation\, we introduce t he concept of a tactical storage system (TSS) that allows users to create\ , reconfigure\, and destroy distributed storage systems without special pr ivileges or complex configuration.\n\nWe have deployed a prototype TSS of 200 disks and 8 TB of storage at the University of Notre Dame and applied it to several problems in astrophysics\, high energy physics\, and bioinfo rmatics. This talk will focus on novel system structures that support dat a-intensive science.\n\nAbout the speaker:\nDouglas Thain is an Assistant Professor of Computer Science and Engineering at the University of Notre D ame. He received a Ph.D. in Computer Science at the University of Wiscons in where he contributed to the Condor distributed computing project. His a reas of research include distributed file systems\, distributed computing\ , fault tolerance\, and mobile computing.\n\n\nOrganiser(s): \nMiguel Ange l Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=a057237 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a057237 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Mumbai CHEP'06 Conference DTSTART;VALUE=DATE-TIME:20060303T083000Z DTEND;VALUE=DATE-TIME:20060303T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a061013@cern.ch DESCRIPTION:The speakers will offer the highlights of the\nCHEP'06 Confere nce\nheld at Mumbai/India\, addressing:\nAlan Silverman - the general summ ary\, including of the plenaries\nGavin Mccance - Distributed Event Produc tion and Processing\nMarkus Schulz - Grid Middleware and e-infrastructure operation\nVlado Bahyl - Computing facilities and networking\nBill Tomlin - Software Tools and Information Systems\nIoannis Papadopoulos - Software Components and Libraries\n\nOrganiser(s): \nMiguel Angel Marquina\nComputi ng Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay. py?confId=a061013 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a061013 END:VEVENT BEGIN:VEVENT SUMMARY:The involvement of IT in EU Computing projects DTSTART;VALUE=DATE-TIME:20060313T090000Z DTEND;VALUE=DATE-TIME:20060313T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a061015@cern.ch DESCRIPTION:The seminar will present IT department's involvement in grid p rojects\nco-funded by the European Commission.\n The speakers will giv e an overview of the objectives and structure of\nthe different EU project s where IT is involved\, as well as how they map\nin practice into the dep artment units.\n\nWolfgang von Rüden and Frederic Hemmer - Introduction\n Bob Jones -\nEGEE-II\n Ian Bird -\nInfrastructure and Application\ n(EGEE/LCG\, EUMedGrid\, EUChinaGrid\, Health-e-Child\, BalticGrid\, EELA\ , SEE-GRID2)\n Alberto di Meglio -\nETICS and OMII-Europe\n Francois Fluck iger -\nISSEG\n Erwin Laure -\nICEAGE\nThese presentations will be followe d by some time dedicated to discussion.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=a061015 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a061015 END:VEVENT BEGIN:VEVENT SUMMARY:Serial and Parallel Random Number Generation DTSTART;VALUE=DATE-TIME:20060329T140000Z DTEND;VALUE=DATE-TIME:20060329T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a061571@cern.ch DESCRIPTION:We will look at random number generation from the point-of-vie w of Monte Carlo computations. Thus\, we will examine several serial meth ods of pseudorandom number generation and two different parallelization te chniques. Among the techniques discussed with be "parameterization\," whi ch forms the basis for the Scalable Parallel Random Number Generators (SPR NG) library. SPRNG was developed several years ago by the author\, and ha s become widely used within the international Monte Carlo community. SPRN G is briefly described\, and the lecture ends with a short revue of quasir andom number generation. Quasirandom numbers offer many Monte Carlo appli cations the advantage of superior convergence rates.\n\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://i ndico.cern.ch/conferenceDisplay.py?confId=a061571 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a061571 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Spring 2006 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20060421T073000Z DTEND;VALUE=DATE-TIME:20060421T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a062002@cern.ch DESCRIPTION:\n\nThe \nspring 2006 meeting of HEPiX has been held at CASPUR \, Roma (Italy) from April 3rd to 7th. HEPiX is a global organization of s ervice managers and support staff providing computing facilities for the H igh Energy and Nuclear Physics community.\n\nThe speakers at this joint Co mputing/After-C5 Seminar will present a summary of the subjects addressed during the meeting.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a062002 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a062002 END:VEVENT BEGIN:VEVENT SUMMARY:Scientific Computing and Apple's Intel Transition DTSTART;VALUE=DATE-TIME:20060426T120000Z DTEND;VALUE=DATE-TIME:20060426T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a061242@cern.ch DESCRIPTION:NOTE UNUSUAL TIME\nFor over five years\, Mac OS X has combined the power of UNIX and the ease of the Macintosh for users around the worl d. This January\, Apple shipped Mac OS X for Intel processors for the firs t time. The entire Macintosh product line will transition to Intel process ors by the end of 2006.\n\nThe presentation will address:\nThe challenges of migrating an entire operating system to a new architecture\nHow the Int el transition affects scientific computing on Mac OS X\nHow to make your o wn software run natively on both PowerPC and Intel-based Macs\nRosetta\, t he translation engine which enables existing PowerPC software to run on In tel-based Macs\nIntel's published processor roadmap and how it may affect the future of personal and scientific computing\nAbout the speaker:\n \n Eric Albert is Senior Software Engineer in Apple's Core Technologies group . During Mac OS X's transition to Intel processors he has worked on almost every part of the operating system\, from the OS kernel and compiler tool s to applications like iPhoto and Keynote. Prior to working on bringing M ac OS X to Intel\, he worked for on the shared-source .NET project ('Rotor ') at Microsoft and for Apple's Java team.\n\n\n(*) In order to assess bes t the size of the room where the seminar will be\nheld\, we kindly ask you to register your (free) attendance\nhere\n\n\nOrganiser(s): \nMiguel Ange l Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=a061242 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a061242 END:VEVENT BEGIN:VEVENT SUMMARY:Public Resource Computing projects at Extremadura DTSTART;VALUE=DATE-TIME:20060912T140000Z DTEND;VALUE=DATE-TIME:20060912T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a063365@cern.ch DESCRIPTION:\nCERN and the regional government of Extremadura/Spain signed a framework\nprotocol of collaboration in February 2005. Since then\, bot h institutions\ncollaborate in several IT projects.\n\nOne of them is the use of PRC technologies (in particular\, BOINC) to make\navailable to Extr emadura and CERN researchers local unused computing\ncapacity. Indeed\, th e regional secondary schools host some 80.000 computers\n(running a Linux/ Debian flavour called "Linex") with abundant\nspare CPU cycles. PRC is hel ping to put such large potential at the disposal\nof a number of scientifi c projects.\n\nThis talk will present the current status of the work carri ed out so far\,\nand the plans to exploit these resources for e-Science ac tivities.\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\n IT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=a0633 65 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a063365 END:VEVENT BEGIN:VEVENT SUMMARY:Web 2.0 and the Semantic Web: Implications for Future HEP Web Appl ications DTSTART;VALUE=DATE-TIME:20060919T140000Z DTEND;VALUE=DATE-TIME:20060919T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a063366@cern.ch DESCRIPTION:Web 2.0 and the Semantic Web represent two of the interesting directions in which Web technology is evolving. Web 2.0 focusses on a 'use r-centric' Web with strong emphasis on community and re-use while the Sema ntic Web presents a rigorous 'data-centric' model. In this talk Bebo will describe these technologies and speculate on how they might be utilized in high-energy physics Web applications and services. Specific strategies an d implementation models will be provided and discussed.\n\nAbout the Speak er:\nBebo White is a Departmental Associate (retired) at SLAC and has spen t considerable time at CERN. In addition\, he holds faculty appointments a t Hong Kong University\, the University of San Francisco\, and Contra Cost a College. He is a frequent speaker at conferences\, academic institutions \, and for commercial organizations around the world. Bebo has been a memb er of the International World Wide Web Conference Committee (IW3C2) since 1996 and in that time has served as General Co-Chair of two of the confere nces and Tutorial and Workshops Co-Chair for four of the conferences. He i s often found on the program committees of the international conferences o n Web Technology and is one of the managing editors of 'The Journal of Web Engineering' and 'The Journal of Computers in Mathematics and Science Tea ching.' Additional information can be found\nhere.\n\n\nOrganiser(s): \nMi guel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indic o.cern.ch/conferenceDisplay.py?confId=a063366 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a063366 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Autumn 2006 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20061020T073000Z DTEND;VALUE=DATE-TIME:20061020T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a063458@cern.ch DESCRIPTION:The Autumn 2006 meeting of HEPiX is being held at Jefferson La b\, USA from October 9th to 13th.\nHEPiX\nis a global organization of serv ice managers and support staff providing computing facilities for the High Energy and Nuclear Physics community.\n\nThe speakers at this joint Compu ting/After-C5 Seminar will present a summary of the subjects addressed dur ing the meeting.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Sem inars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?con fId=a063458 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a063458 END:VEVENT BEGIN:VEVENT SUMMARY:Evolution of clustered storage DTSTART;VALUE=DATE-TIME:20070308T150000Z DTEND;VALUE=DATE-TIME:20070308T170000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a0769@cern.ch DESCRIPTION:The session actually featured two presentations:\nEvolution of clustered storage\nby Lance Hukill\, Quantum Corporation\nALICE DAQ - Usa ge of a Cluster-File System: Quantum StorNext\nby Pierre Vande Vyvre\, CER N-PH\n\nthe second one prepared at short notice by Pierre (thanks!) to pre sent how the Quantum technologies are being used in the ALICE experiment. The abstract to Mr Hukill's follows.\nClustered Storage is a technology th at is driven by business and\nmission applications. The evolution of Clust ered Storage solutions\nstarts first at the alignment between End-users ne eds and Industry\ntrends:\nPush-and-Pull between managing for today versus planning for tomorrow\nBreaking down the real business problems to the co re applications\nCommoditization of clients\, servers\, and target devices \nInterchangeability\, Interoperability\, Remote Access\, Centralized\ncon trol\nOh\, and yes\, there is a budget and the "real world" to deal with\n This presentation will talk through these needs and trends\, and then ask\ nthe question\, "are we aligned"?\nAbout the speaker\n\nAs Vice President of\nQuantum's Worldwide Software Sales\, Lance Hukill\nhas responsibility for Software Sales and Marketing serving new and\nglobal markets outside o f Backup & Recovery\, which primarily encompasses\nFile Serving & Archive. This role includes the management of OEM and\nAlliance relationships\, as well as Branded channel and partnerships\nsupporting Vertical and Geograp hic driven go-to-market strategies and\nexecution.\n\nJoining Quantum (via \nADIC)\nin 1999\, Mr. Hukill previously served as\nExecutive Director of the Advanced Solutions Business Unit with\nresponsibility for Sales\, Mark eting\, and Product Marketing and\nManagement. Prior to that\, Mr. Hukill served as an Account Executive and\nlater Area Sales Director for the Fede ral Government. This includes\norganizational alignment for sales\, channe l\, business development\, and\npre-sales engineering across Defense\, Int elligence\, and Civilian sales\ndistricts.\n\nPrior to ADIC\, Mr. Hukill s pent two years at SANZ (formerly Vangard\nTechnology) with roles as Accoun t Executive and System Engineering\nManager. Additionally\, Mr. Hukill spe nt 6 years with Harris Corporation\nin a variety of Engineering and Manage ment positions supporting programs\nwithin the Intelligence Community.\n\n Mr. Hukill holds a BS in Electrical Engineering from the Colorado School\n of Mines.\n\nNote from the organiser - For the curious\, some press notes on Quantum found via Google:\nAug 22\, 2006 - Quantum completes acquisitio n of Advanced Digital Information Corporation\nOct 26\, 2006 - Quantum shi ps latest version of data management software\nFeb 20\, 2007 - Quantum int egrates DLT-S4 tape drive into scalar libraries\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=a0769 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=a0769 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Spring 2007 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20070504T073000Z DTEND;VALUE=DATE-TIME:20070504T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a0770@cern.ch DESCRIPTION:The Spring 2007 meeting of HEPiX is being held at DESY\, Hambu rg from April 23rd to 27th.\nHEPiX\nis a global organization of service ma nagers and support staff providing computing facilities for the High Energ y and Nuclear Physics community.\n\nThe speakers at this joint Computing/A fter-C5 Seminar will present a summary of the subjects addressed during th e meeting.\n\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminar s /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId= a0770 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a0770 END:VEVENT BEGIN:VEVENT SUMMARY:The RapidMind Development Platform for Cell\, GP-GPU\, and Multico re CPUs DTSTART;VALUE=DATE-TIME:20070510T120000Z DTEND;VALUE=DATE-TIME:20070510T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a0792@cern.ch DESCRIPTION:\nMassively multi-core processors such as GPUs and the Cell BE have the\npotential to deliver high performance computation to many appli cations.\nHowever\, these processors require parallel programming on sever al\nlevels\, use some novel programming models\, and native code written f or\none will not execute on the other. The\nRapidMind\ndevelopment platfo rm\nenables portable access to the power of these processors. It provides a\nuniform\, simple\, safe\, data-parallel programming model and takes ca re of\nmost of the low-level details of mapping programs to each hardware\ ntarget. It combines a dynamic compiler with a runtime streaming\nexecutio n manager\, and provides a single system image for computations\nrunning o n any number of cores.\n\nThe interface to this system is embedded inside ISO standard C++\, where\nit can capture arbitrary computations specified at runtime. The use of\na dynamic compiler means that high-level C++ abst ractions can be used\nwithout sacrificing performance\, while maintaining portability among\ncurrent--and future--processors.\n\nThis seminar will i ntroduce and demonstrate the use of the RapidMind\nplatform on GPUs (for b oth visualization and general-purpose\ncomputation)\, the Cell BE (where i t runs on both IBM blades and the PS3)\nand on multicore CPU (x86 processo r targets are under development).\nComparative performance\nresults will b e presented for the GPU\, the Cell BE\, and multicore CPUs.\n\nOn\nGPUs up to a 30x speedup over tuned CPU code has been achieved. On the\nCell\, we can match or exceed the performance of vendor tools in tuned\napplicati ons\, but with much less developer effort. On CPUs\, we have\ndoubled perf ormance over native tools (i.e. 8x speedup on a 4-core\nrelative to icc on one).\n\nAbout the speaker\n\nMichael McCool is an Associate Professor in the School of Computer\nScience at the University of Waterloo and co-foun der and Chief Scientist\nof RapidMind Inc.\n\nProf. McCool's current resea rch efforts are targeted at enabling\nhigh-performance parallel applicatio ns by the development of advanced\nprogramming technologies. Research int erests include interval analysis\,\nMonte Carlo and quasi-Monte Carlo nume rical methods\, optimization\,\nsimulation\, sampling\, cellular automata\ , real-time computer graphics\,\nvision\, image processing\, hardware desi gn\, and programming languages and\ndevelopment platforms.\n\nHe has degre es in both Computer Engineering (B.A.Sc. with Math option\,\nWaterloo\, 19 89\, Sir Sandford Medal) and Computer Science (M.Sc. in 1991\nand Ph.D. in 1995\, Toronto).\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Semi nars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?conf Id=a0792 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a0792 END:VEVENT BEGIN:VEVENT SUMMARY:Accelerating data communications to produce a tenfold performance improvement for Office/Database/Web/Media applications over the WAN/Grid DTSTART;VALUE=DATE-TIME:20070511T120000Z DTEND;VALUE=DATE-TIME:20070511T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a07104@cern.ch DESCRIPTION:\nProf Wang and his group have developed a data communication\ nprotocol\, which beats classic ones by a factor of 25 over a real-world l ink\nof 10\,000 km. This protocol is the first of its kind worldwide. Best of all\,\nthis protocol requires no changes in the way users work with th eir\napplications since it conforms the existing IT infrastructures. Durin g his\npresentation\, he is going to show videos of using this protocol to \naccelerate applications\, ranging from Office\, Database\, Web Browser t o Media\nPlayer.\n\nAbout the speaker\n\nProf. Frank Wang is a Professor a nd Chair in e-Science and Grid Computing\,\nDirector of Centre for Grid Co mputing within the context of\nCambridge-Cranfield High Performance Comput ing Facility (CCHPCF).\nProf Wang's appointment is seen\nas crucial to the initiative of the CCHPCF\, which is a collaborative\nresearch facility in the Universities of Cambridge and Cranfield. Prof Wang\nhas a publication record including a book titled "Encyclopedia of Grid\nComputing" (underwa y) and 57 journal papers [12 IEEE Transactions\, 5 JAP\, 1\nAPL\, 2 ACM pu blications\, etc]. Prof Wang is on the Editorial Board of 4\ninternational journals. He serves the High End Computing Panel for Science\nFoundation Ireland (SFI). He has been elected as the Chairman (UK & Republic\nof Irel and Chapter) of the IEEE Computer Society from January 2005. He has\nbeen invited to give talks at Princeton University\, Cambridge University\n(e-S cience Centre)\, Oxford University\, Birmingham University\, Manchester\nU niversity\, etc.\n\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing S eminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=a07104 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a07104 END:VEVENT BEGIN:VEVENT SUMMARY:The future of Information Search DTSTART;VALUE=DATE-TIME:20070523T090000Z DTEND;VALUE=DATE-TIME:20070523T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a07116@cern.ch DESCRIPTION:\nSearch technology is fundamental in transforming ever increa sing amounts\nof internal and external data into useful information. The a mount of\nprimary\, scientific and administrative data available to resear chers\,\nadministrators and analysts creates environments characterized by information overload.\n\nCERN is presently implementing a search solution based on the\nproduct of the Norwegian company FAST\, one of the leading companies\nin search technologies. As an introduction\, a brief overview o f the\nCERN Search service\, that is based on the FAST ESP solution\, will be\ngiven.\n\nIn the main part of the presentation\, Torstein Thorsen\, w ill present\nscenarios\, concepts and developments in the area of informat ion access\ntechnology. He will give an insight in the evolution and the f uture of\nsearch technology. Furthermore\, he will address aspects of info rmation\ndiscovery and metadata extraction for scientific documents\, data \ncleansing\, ad hoc data analysis and visualization\, and processing and search\nsolutions for complex structured data.\n\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.ce rn.ch/conferenceDisplay.py?confId=a07116 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=a07116 END:VEVENT BEGIN:VEVENT SUMMARY:The Dynamics of Information DTSTART;VALUE=DATE-TIME:20070612T090000Z DTEND;VALUE=DATE-TIME:20070612T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-a07119@cern.ch DESCRIPTION:\nThe dynamics of information within organizations is relevant to issues\nof productivity\, innovation\, and the sorting out of useful i deas from\nthe general chatter of a community. How information spreads and is\naggregated determines the speed with which individuals and organizati ons\ncan act and plan their future activities.\n\nThis talk will describe mechanisms for automatically identifying\ncommunities of practice within o rganizations and for elucidating\nthe spread of information spreads within those communities. In addition\,\na new method for forecasting uncertain events using small groups of\npeople will be presented\, along with empiri cal results that show its\nefficacy at making predictions in the real worl d.\n\nAbout the speaker\n\nBernardo Huberman is a Senior HP Fellow and Dir ector of the Information Dynamics Lab at Hewlett Packard Laboratories. He received his Ph.D. in Physics from the University of Pennsylvania\, and is currently a Consulting Professor in the Department of Applied Physics at Stanford University. He originally worked in condensed matter physics\, ra nging from superionic conductors to two-dimensional superfluids\, and made contributions to the theory of critical phenomena in low dimensional syst ems. He was one of the discoverers of chaos in a number of physical system s\, and also established a number of universal properties in nonlinear dyn amical systems. His research into the dynamics of complex structures led t o his discovery of ultradiffusion in hierarchical systems.\n\nYou may read more about him\nhere\n\nOrganiser(s): \nMiguel Angel Marquina\nComputing Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=a07119 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=a07119 END:VEVENT BEGIN:VEVENT SUMMARY:Ethics and Practice of Free Software DTSTART;VALUE=DATE-TIME:20070618T150000Z DTEND;VALUE=DATE-TIME:20070618T170000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-17556@cern.ch DESCRIPTION:Dr. Richard Stallman is coming to Switzerland\non June 18-19 t o make the keynote\nspeech of the conference\n\n\nFree and open source sof tware: challenges and opportunities\n\nto be held at Lausanne.\nHe is our guest in the afternoon of June 18th. \n\nAbout the speaker\n\nRichard Matt hew Stallman is a software freedom activist\,\nhacker\, and software devel oper. In September 1983\, he launched the\nGNU Project to create a free Un ix-like operating system\, and has been\nthe project's lead architect and organizer. With the launch of the\nGNU project he started the free softwar e movement\, and in October\n1985 set up the Free Software Foundation. He co-founded the League\nfor Programming Freedom. Stallman pioneered the con cept of copyleft\nand is the main author of several copyleft licenses incl uding the GNU\nGeneral Public License\, the most widely used free software license.\n(from Wikipedia)\n\nOrganiser(s): \nMiguel Angel Marquina\nComp uting Seminars /\nIT Department\n\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=17556 LOCATION:CERN SALLE DU CONSEIL URL:http://indico.cern.ch/conferenceDisplay.py?confId=17556 END:VEVENT BEGIN:VEVENT SUMMARY:Applications of Evolutionary Algorithms in High Energy Physics and Computing DTSTART;VALUE=DATE-TIME:20070720T130000Z DTEND;VALUE=DATE-TIME:20070720T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-17771@cern.ch DESCRIPTION:Evolutionary algorithms are computer algorithms inspired by bi ological evolutionary theories. They are commonly used for solving benchma rk problems in computer science as well as complex real-world problems. In high energy physics and the related fields these algorithms have been suc cessfully tested and their presence gradually increases. This seminar will review classical evolutionary algorithms such as Genetic Algorithms\, Gen etic Programming and Evolutionary Strategies\, and will introduce new deve lopments in the field such as Gene Expression Programming. Applications o f these algorithms for solving high energy physics data analysis problems and computing related tasks such as job scheduling optimisation on computa tional grids will be reviewed and discussed.\n\nAbout the speaker\nLiliana Teodorescu is a Lecturer at Brunel University\, London\, UK. She holds a Ph.D. in Particle Physics and has more than 10 years research experience o n medium and large scale particle physics experiments. She worked at renow n laboratories around the world: Thomas Jefferson National Accelerator Fac ility (TJNAF)\, USA\, Instituto Nazionale di Fisica Nucleare (INFN) – Pi sa\, Italy\, Stanford Linear Accelerator Centre (SLAC)\, USA and European Laboratory for Particle Physics (CERN)\, Geneva. She currently works on C MS (CERN) and BaBar (SLAC) experiments. In addition of physics research in these experiments\, she is interested in cross-disciplinary research invo lving particle physics and various fields of computer science for developm ent of novel data analysis techniques for particle physics.\n\n\nhttp://in dico.cern.ch/conferenceDisplay.py?confId=17771 LOCATION:CERN IT AMPHITHEATRE URL:http://indico.cern.ch/conferenceDisplay.py?confId=17771 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Victoria BC CHEP'07 Conference DTSTART;VALUE=DATE-TIME:20070921T073000Z DTEND;VALUE=DATE-TIME:20070921T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-18484@cern.ch DESCRIPTION:This year's Computing in High Energy and Nuclear Physics (CHEP ) conference will be held in Victoria\, British Columbia\, Canada from 2-7 September 2007. A WLCG Meeting will be held on from 1-2 September prior t o CHEP 2007.\n\nCHEP conferences provide an international forum to exchang e information on computing experience and needs for the High Energy Physic s and Nuclear Physics communities\, and to review recent\, ongoing and fut ure activities.\n\nThe speakers at this Computing Seminar will present the ir views of the CHEP'07 conference.\n\n\nhttp://indico.cern.ch/conferenceD isplay.py?confId=18484 LOCATION:CERN IT AMPHITHEATRE URL:http://indico.cern.ch/conferenceDisplay.py?confId=18484 END:VEVENT BEGIN:VEVENT SUMMARY:Programming Challenges for Many-Core Computing DTSTART;VALUE=DATE-TIME:20071011T073000Z DTEND;VALUE=DATE-TIME:20071011T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-22280@cern.ch DESCRIPTION:Many-core architectures face significant hurdles to successful \nadoption by ISVs\, and ultimately\, the marketplace. One of the most\ndi fficult is addressing the programmability problems associated\nwith parall el computing. For example\, it is notoriously difficult\nto debug a parall el application\, given the potential interleavings\nof the various threads of control in that application. Another\nproblem is that predicting perfo rmance\, even at coarse accuracy\,\nis extremely inaccurate. I will explai n why a chip company like\nIntel is interested in advanced programming lan guages research and\nbelieves this is critical to adoption of many-core ar chitectures.\n\nIntel's Programming Research Lab is addressing these issue s for\nboth client and server computing\, in particular media and gaming\n workloads. We are implementing a high-level programming\nabstractions base d on transactional memory\, data parallel\nprogramming models and function al languages. In this talk\, I will\nbriefly discuss a language based on N ested Data Parallelism (NDP)\ncalled Ct. NDP models have the advantage of being deterministic\,\nmeaning that the functional behaviors of sequential and parallel\nexecutions of an NDP program are always the same for the sa me\ninput. Data races are not possible in this model. Furthermore\,\nNDP m odels have an easy to understand coarse performance model\,\nwhich can be made more accurate for specific architectural\nfamilies. This enables the programmer to comprehend the\nperformance implications of their code well- enough to make well-\ninformed algorithmic choices.\n\nAbout the speaker\n Anwar Ghuloum earned degrees at the University of California\, Los\nAngele s (B.S.\, Computer Science and Engineering) and Carnegie\nMellon Universit y's School of Computer Science (Ph.D.\, Computer\nScience\, 1996)\, where his thesis introduced concepts of Nested\nData Parallel idioms to traditio nal parallelizing compilers. Anwar\nhas been a Senior Staff Scientist with Intel's Programming Systems\nLab since joining in early 2002\, working on diverse topics such as\noptimizing memory system performance\, parallel a rchitecture\nevaluation\, parallel language and compiler design\, and mult imedia\napplications.\n\nBefore that\, he co-founded and was the CTO of a fab-\nless semiconductor startup called Intensys that built\nprogrammable\ , highly parallel image and video processors for the\nconsumer electronics market. Prior to that\, Anwar developed novel\npredictive drug design sof tware for early lead optimization using\n3D surface pattern recognition te chniques for a biotech startup\ncalled MetaXen (acquired by Exelexis Pharm aceuticals). He has\nalso served as a post-doctoral research associate at Stanford\nUniversity's Computer Science department. A recurring theme in\n Anwar's work has been to bridge high-level application knowledge\nand low- level parallel architecture constraints with careful\nparallel language an d compiler design.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=22 280 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=22280 END:VEVENT BEGIN:VEVENT SUMMARY:Managing Virtual Machine Images DTSTART;VALUE=DATE-TIME:20071102T083000Z DTEND;VALUE=DATE-TIME:20071102T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-20468@cern.ch DESCRIPTION:\nHardware virtualization allows dynamic deployment of differe nt operating\nsystem flavours across different computers. It also allows i ncreased\nflexibility in terms of tayloring OSs. A grid job or service can have its\nsoftware dependencies satisfied in a VM image\, and this can be shipped with\nthe job or service as a complete execution environment. The openlab project\,\n"OSFarm\," aims to dynamically generate different Linu x flavours and complete\nself-contained services\, also called "virtual ap pliances\," and maintain them\nin a repository. VM images can be several h undreds of MB large\, so\ntransferring them from a repository to a target machine can take a\nconsiderable amount of time and congest the network. A second project\,\n"Content Based Transfer\," uses hashes to identify iden tical blocks between\nimages in order to speed up transfer and reduce netw ork congestion. The talk\nwill discuss the designs behind these two proje cts and show scenarios of\nusage.\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=20468 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=20468 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Fall 2007 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20071130T083000Z DTEND;VALUE=DATE-TIME:20071130T103000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-23512@cern.ch DESCRIPTION:\nThe\nFall 2007 meeting of HEPiX is being held at Washington University in St. Louis\n\, USA\, from November 5th to 9th.\n\n\nHEPiX\nis a global organization of service managers and support staff providing com puting facilities for the High Energy and Nuclear Physics community. \n\nT he speakers at this joint Computing/After-C5 Seminar will present a summar y of the subjects addressed during the meeting. \n\n\nhttp://indico.cern.c h/conferenceDisplay.py?confId=23512 LOCATION:CERN 513-1-024 URL:http://indico.cern.ch/conferenceDisplay.py?confId=23512 END:VEVENT BEGIN:VEVENT SUMMARY:Debugging and Optimizing Applications for Multi core MPP Architect ures DTSTART;VALUE=DATE-TIME:20071219T100000Z DTEND;VALUE=DATE-TIME:20071219T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-24297@cern.ch DESCRIPTION:\nAs two\, four and potentially eight-core processors become t he norm\, the\nde facto HPC architecture is tending towards large clusters of modest 8-16\ncore shared-memory servers\, potentially with co-processi ng devices (e.g..\nGPGPUs\, FPGAs\, Clearspeed). Programming these machine s optimally presents a\nnumber of challenges\, and applications that use m ixed programming models are\nnow becoming commonplace.\n\n\nIn this presen tation we will discuss the challenges facing today's HPC\napplication deve lopers\, and the need for simple software tools that can\naddress mixed pr ogramming models (MPI and OpenMP/threads). We will present\nnew multicore features of Allinea's Distributed Debugging Tool (DDT) and\nOptimisation a nd Profiling Tool (OPT)\, and discuss our aims to provide a\nconsolidated\ , scalable\, yet intuitive framework for HPC developers.\n\n\nWe will also discuss the competitive software tools market and compare arch rival prod ucts with both DDT and OPT and show where the Allinea software\ntools brin g greatest benefit.\n\nThis will be followed by a demo of both DDT and OPT and a Q&A session at the\nend.\n\nhttp://indico.cern.ch/conferenceDisplay .py?confId=24297 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=24297 END:VEVENT BEGIN:VEVENT SUMMARY:Event reconstruction in HEP experiments DTSTART;VALUE=DATE-TIME:20080207T150000Z DTEND;VALUE=DATE-TIME:20080207T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-28013@cern.ch DESCRIPTION:In modern high energy physics experiments one has to process t erabytes of\ndata at 10 MHz input rates. Therefore\, keeping the same qual ity of event\nreconstruction the algorithms for event reconstruction have to be speeded up\nsignificantly in order to be used on-line. The most wide ly used methods for\nevent reconstruction are presented and special featur es of their\nimplementation are discussed.\n\nhttp://indico.cern.ch/confer enceDisplay.py?confId=28013 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=28013 END:VEVENT BEGIN:VEVENT SUMMARY:Scalla Advancements DTSTART;VALUE=DATE-TIME:20080509T120000Z DTEND;VALUE=DATE-TIME:20080509T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-33542@cern.ch DESCRIPTION:\nThis talk will discuss new developments\, enhancements\, and \ndirections in xrootd clustering. Included will be presentations of the\n xrootd file system\, creating a composite cluster name space\, SRM\nintegr ation\, cluster globalization\, and more.\n\nhttp://indico.cern.ch/confere nceDisplay.py?confId=33542 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=33542 END:VEVENT BEGIN:VEVENT SUMMARY:Creating Secure Software DTSTART;VALUE=DATE-TIME:20080515T120000Z DTEND;VALUE=DATE-TIME:20080515T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-33531@cern.ch DESCRIPTION:\nComputer security has been an increasing concern for IT prof essionals for\na number of years\, yet despite all the efforts\, computer systems and\nnetworks remain highly vulnerable to attacks of different kin ds. Design\nflaws and security bugs in the underlying software are among t he main\nreasons for this situation.\n\nThis lecture addresses the followi ng question: how to create secure\nsoftware? The lecture starts with a def inition of computer security and an\nexplanation of why it is so difficult to achieve. It then introduces the\nmain security principles (like least- privilege\, or defense-in-depth) and\ndiscusses security in different phas es of the software development cycle.\nThe emphasis is put on the implemen tation part: most common pitfalls and\nsecurity bugs are listed\, followed by advice on best practice for security\ndevelopment.\n\n\nhttp://indico. cern.ch/conferenceDisplay.py?confId=33531 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=33531 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Spring 2008 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20080523T073000Z DTEND;VALUE=DATE-TIME:20080523T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-21328@cern.ch DESCRIPTION:\nThe Spring 2008 meeting of HEPiX is being held at CERN on 5- 9 May 2008. HEPiX is a global organization of service managers and support staff providing computing facilities for the High Energy and Nuclear Phys ics community. \n\n\nThe speakers at this joint Computing/After-C5 Seminar will present a summary of the subjects addressed during the meeting:\nGen eral highlights - Alan Silverman\nScientific Linux - Tim Bell\nComputer Ce ntre issues - Tony Cass\nBenchmarking - Alex Iribarren\nStorage - Andreas Peters\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=21328 LOCATION:CERN URL:http://indico.cern.ch/conferenceDisplay.py?confId=21328 END:VEVENT BEGIN:VEVENT SUMMARY:Microsoft's HPC and cluster strategy DTSTART;VALUE=DATE-TIME:20080616T133000Z DTEND;VALUE=DATE-TIME:20080616T143000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-35369@cern.ch DESCRIPTION:\nHPC has undergone a revolution with the advent of commodity compute\nclusters. Where once HPC was the domain of a specialized few rese archers\,\nnow virtually any enterprise can gain access to large aggregate s of\nprocessing power - holding out the promise of truly "personal\nsuper computing." Personal supercomputing will enable scientists and\nengineers to interactively employ massive computational and data processing\nresourc es to solve complex technical problems. They will be able to assemble\ncom plex models from best of breed scientific applications and most recent\nda ta sources that might span organizations and institutions\, while\nintegra ting with the collaborative and business workflows. To make this\nrevoluti on complete we need to reach out to organizations and users with an\nenvir onment that is familiar and easy to use and exploit. This is where\nMicros oft products\, user interfaces\, and developer tools add value\, by\ninteg rating these specialized resources into an environment that allows\nseamle ss computing\, communication\, collaboration and ultimately creativity\nan d productivity.\n\nAbout the speaker\n\nKyril Faenov is an entrepreneur\, technologist and executive with a 16 year\ntrack record of global impact i n the high technology industry. Presently\,\nKyril is a General Manager a t Microsoft Corporation\, responsible for one of\nthe top Windows Server g rowth initiatives and bringing High Performance\nComputing into mainstream adoption. Kyril founded the HPC business unit in\n2004 and lead the devel opment of Windows Compute Cluster Server product\,\nsales and marketing\, and world-wide industry and academic partnerships. His\nprior contribution s at Microsoft include running the planning process for\nWindows Server 20 08\, co-founding a distributed systems project in the office\nof the CTO t hat later became the Robotics SDK\, driving cross-company\nstrategic initi atives as a technical advisor to the Platforms Group Vice\nPresident\, and developing the Network Load Balancing scale-out technology in\nWindows 20 00. Kyril joined Microsoft in 1998 as the result of acquisition of\nValenc e Research\, an Internet server clustering startup he co-founded and\ngrew to profitability by securing MSN\, Microsoft.com and some of the world's\ nother largest web sites as its clients. Prior to that\, he was a principa l\ntechnical member of two other parallel systems startups\, as well as a\ nsoftware consultant to Intel on a variety of embedded and performance\nsi mulation projects. Kyril holds BS and MS degrees in computer science and\n an MBA in Technology Management. Kyril grew up in Moscow\, Russia\, where he\ntrained with the Olympic Reserve swimming team and studied physics and \nmathematics.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=3536 9 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=35369 END:VEVENT BEGIN:VEVENT SUMMARY:Wuala - a distributed file system DTSTART;VALUE=DATE-TIME:20080617T120000Z DTEND;VALUE=DATE-TIME:20080617T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-31399@cern.ch DESCRIPTION:\nWuala\nis a new way of storing\, sharing\, and publishing fi les on the\ninternet. Unlike traditional online storage systems\, Wuala is decentralized\nand can harness idle resources of participating computers to build a large\,\nsecure\, and reliable online storage. This enables its users to trade parts\nof their idle local disk space for online storage a nd it allows us to\nprovide a better service for free. We started research and development on\nWuala in fall 2004 at the Swiss Federal Institute of Technology (ETH Zurich)\nand we founded our company in 2007. A couple of m onths ago\, we launched the\nclosed alpha version of Wuala\, which has bee n a great success. This summer\,\nwe will launch the public beta which eve rybody can join. In the talk\, I will\nexplain what Wuala is and how it wo rks.\nAbout the speaker\n\nDominik Grolimund\n(dominik@wua.la)\nis 27 year s old and has studied computer science at ETH\nZurich. In 1998\, he founde d his software company Caleido\, and developed the\nCaleido Address-Book\, a professional contact management software\, of which\nover 35'000 licens es have been sold so far in Switzerland\, Germany and\nAustria.\n\nIn 2003 \, he did an exchange semester at the TU Delft\, the Netherlands\, as\npar t of the Unitech exchange program\, focusing on business and management.\n In 2004\, a six-month internship followed with Siemens Corporate Research in\nPrinceton\, New Jersey in the US\, where he worked in the 'Intelligent Vision\n& Reasoning' department\, developing a product for power plant mo nitoring.\n\n\nIn 2004\, he started project Wuala at ETH Zurich together w ith Luzius Meisser and in 2007\, they founded their company to further dev elop Wuala and launch it publicly.\n\nHe is open to technical exchanges on this project\, feel free to contact him.\n\n\nhttp://indico.cern.ch/confe renceDisplay.py?confId=31399 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=31399 END:VEVENT BEGIN:VEVENT SUMMARY:Building PetaFLOP HPC systems with energy-efficient processors DTSTART;VALUE=DATE-TIME:20080707T120000Z DTEND;VALUE=DATE-TIME:20080707T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-36588@cern.ch DESCRIPTION:\nGodson-3 is a multi-core processor based on the 64-bit super scalar Godson-2 CPU core\, which in turn is a four-issue\, out-of-order ex ecution CPU which runs the MIPS64 instruction set.\nThe CPU core of Godson -3 is enhanced to support efficient X86 to MIPS binary translation\, and t o optimize performance\, power consumption\, reliability and debug methods .\nGodson-3 adopts two-dimension mesh topology where each node in the mesh includes an 8*8 crossbar which connects four processor cores.\nThe low po wer design makes the 4-core Godson-3 consume 10 watts at 1GHz. It will be taped out in first half of 2008.\n\n\nOne important application of Godson- 3 is the low cost high performance\ncomputing (HPC). Based on Godson-3\, t he design of a national PetaFLOP HPC system and a personal TeraFLOP HPC sy stem are in plan. This presentation will introduce the HPC plans based on the Godson-3 multi-core processor.\n\n\nhttp://indico.cern.ch/conferenceDi splay.py?confId=36588 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=36588 END:VEVENT BEGIN:VEVENT SUMMARY:Data Centers Now and in the Future: High Powered and Green! DTSTART;VALUE=DATE-TIME:20080807T090000Z DTEND;VALUE=DATE-TIME:20080807T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-38857@cern.ch DESCRIPTION:\nData Centers today are a close attempt to build for a need t hat is\nonly as far as 2 years out. The trending for high density\nenviron ments is causing a paradigm shift in the way Data Center\ndesigners and Co location Vendors are implementing the next generation\nsupport infrastruct ure for the Internet. This presentation will\nreview power and cooling tec hnologies available today and in the\nfuture to support the Internet. Addi tionally a 15 minute Google Earth\ntour of Large Internet Exchange Points will explain the reasoning for\nchoosing locations in North America and Eu rope.\n\nAbout the speaker\n\nJosh Snowhorn is a Vice President at\nTerrem ark Worldwide\nresponsible for\nall Peering related services for the compa ny. He has been with Terremark\nsince the inception of the NAP (Network Ac cess Point) of the Americas\n(NOTA)\, the largest carrier neutral Data Cen ter in the Western\nhemisphere\, including being involved with the site se lection\, design and\nconstruction of the facility. He has been a member o f the global Peering\ncommunity for many years including being the founder of the Global\nPeering Forum events.\n\n\nhttp://indico.cern.ch/conferenc eDisplay.py?confId=38857 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=38857 END:VEVENT BEGIN:VEVENT SUMMARY:XRootd explained DTSTART;VALUE=DATE-TIME:20080903T090000Z DTEND;VALUE=DATE-TIME:20080903T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-38458@cern.ch DESCRIPTION:\nThis talk gives a broad view of the Scalla/XRootd platform f or data access\nand storage\, intended for a general audience. Emphasis is given to\nexplaining the overall vision about the project (formerly born as a\ncollaboration between SLAC and INFN\, but now treated as a generic o pen\nsource project\, with several contributors\, CERN being among the fir st ones)\, in the form of a short glance at the key features.\n\n\n Althou gh some of the topics have highly technical implications\, we don't\nfocus directly on them\, preferring to give a view about answers to questions\n like "what does it do"\, "how does it work"\, "why was it done that way"\, or\n"how can I set up a simple instance with the latest features".\n\n\nh ttp://indico.cern.ch/conferenceDisplay.py?confId=38458 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=38458 END:VEVENT BEGIN:VEVENT SUMMARY:Using Linux as Hypervisor with KVM DTSTART;VALUE=DATE-TIME:20080915T090000Z DTEND;VALUE=DATE-TIME:20080915T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-39755@cern.ch DESCRIPTION:\nThe presentation will be about the KVM design\, how it is di fferent from\nother virtualization designs\, how shadow pagetables are man aged by\nKVM\, how mmu notifiers can allow paging the guest physical memor y\,\nand a short overview on the paravirtualization and pci-passthrough\nc apabilities of KVM.\n\nAbout the speaker\n\nAndrea Arcangeli joined\nQumra net\nin 2008 because of his interest in\nworking on the KVM virtualization product\, with special interest in\nthe virtual machine memory management . Before joining Qumranet he\nworked for Novell/SUSE since 1999 on many pa rts of the Linux Kernel\,\nbut especially on the virtual memory subsystem of the kernel. He\nstarted working with Linux in his spare time shortly af ter he\nconnected the first time to the Internet back in 1996 while studyi ng\nat Bologna University. He enjoys spending most of his time solving\nso ftware problems\, and trying to promote the adoption of Linux and\nOpen So urce software everywhere.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=39755 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=39755 END:VEVENT BEGIN:VEVENT SUMMARY:Control systems in practice - Safety issues in modern railway traf fic control DTSTART;VALUE=DATE-TIME:20080919T083000Z DTEND;VALUE=DATE-TIME:20080919T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-40759@cern.ch DESCRIPTION:\nAnsaldo STS\nhas developed architectures for electronic and computer based platforms devoted to railways signalling applications\, sin ce the last 70’s.\nLater CENELEC standards have been produced and issued \, and have become the mandatory regulations\, both in Europe and worldwid e\; Ansaldo has adopted the CENELEC “context” and integrated those tec hniques\, where needed\, to make those architectures fully compliant with the CENELEC SIL4 requirements.\n\n\nAnsaldo STS staff is visiting CERN an d offer a half-day event on their activities\, starting with a general pre sentation during the morning. The afternoon will be dedicated to more spec ific presentations although they are equally open to the public.\n\n\nhttp ://indico.cern.ch/conferenceDisplay.py?confId=40759 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=40759 END:VEVENT BEGIN:VEVENT SUMMARY:Inside RAC DTSTART;VALUE=DATE-TIME:20081112T130000Z DTEND;VALUE=DATE-TIME:20081112T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-44257@cern.ch DESCRIPTION:\nThis seminar presents an alternative view of RAC scalability \nissues by investigating the messages exchanged between\ninstances across the private network (interconnect).\n\n\nFollowing a brief introduction t o RAC concepts we will examine\nhow Global Cache Services (GCS) implements Cache Fusion by\nmanaging block locking and consistency across the cluste r. For\neach of the main GCS wait events we will drill down to the\nindivi dual UDP messages that are exchanged by the instances\ninvolved in the ope ration. We will also look at packet size\,\nSCN synchronization and consid er situations in which\nengineering instance affinity and dynamic resource mastering\ncan be beneficial.\n\n\nFinally we will look at a couple of Or acle 11g enhancements to\nthe Cache Fusion algorithm and evaluate their po tential impact\non performance.\n\nAbout the speaker\n\n Julian Dyke is a n independent consultant specializing in Oracle\ndatabases. He has over 20 years of database experience\nincluding 18 years as an Oracle DBA\, devel oper and consultant.\nHis main areas of interest are Oracle performance\, clustering\nand high availability. He is Chair of the\nUKOUG\nRAC & HA SIG \nand co-author of Pro Oracle 10g RAC on Linux. He also\nmaintains www.ju liandyke.com which specializes in Oracle\ndiagnostics\, optimization and internals.\n\n\nHe regularly presents at conferences\, seminars and user g roup\nmeetings in the UK\, Europe\, USA and Asia..\n\n\n\nhttp://indic o.cern.ch/conferenceDisplay.py?confId=44257 LOCATION:CERN 160-1-009 URL:http://indico.cern.ch/conferenceDisplay.py?confId=44257 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Autumn 2008 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20081121T083000Z DTEND;VALUE=DATE-TIME:20081121T103000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-43997@cern.ch DESCRIPTION:\nThe Autumn 2008 meeting of HEPiX was held \nat Taipei on 20- 24 October 2008.\nHEPiX is a global organization of service managers and s upport staff providing computing facilities for the High Energy and Nuclea r Physics community. \n\n\nThe speakers at this joint Computing/After-C5 S eminar will present a summary of the subjects addressed during the meeting :\nGeneral highlights - Helge Meinhard\nMiguel Coelho Dos Santos\nSteven M urray\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=43997 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=43997 END:VEVENT BEGIN:VEVENT SUMMARY:UNOSAT Grid project: using the grid to handle satellite images DTSTART;VALUE=DATE-TIME:20081126T100000Z DTEND;VALUE=DATE-TIME:20081126T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-44982@cern.ch DESCRIPTION:\nThe UNOSAT Grid Project story began at CERN during the Tsuna mi\ndisaster at the end of 2005. \n\n\nDuring this crisis\, CERN provided its help to UNOSAT\, a United\nNations initiative that provides the human itarian community with\naccess to satellite imagery. UNOSAT and a CERN   transferred a great\namount of data related to the Asian Tsunami into the CASTOR storage\nmanagement system and set a software infrastructure enabli ng users to\naccess this data.\n\n\nFrom this initial collaboration more a ctivity related to the grid\nhave started in a collaboration between CERN and UNOSAT\, together\nwith the Geneva University and the ecole engenieurs also in Geneva. \nIn order to enable UNOSAT operators to have access to their data on a\nfully disctributed way\, a new phase of the collaboration was started.\n\n\nWe present here valuable tools emerged from this collab oration\,\noffering field workers new faciltiy to access their data. The f irst\npart will describe the architecture simplifying the grid complexity\ nto the front-end user and the management of geospatial metadata on\nthe g rid. In the second part we will demonstrate some of the tools\ndesigned fo r the UNOSAT field workers : the web and the mobile grid\nportal access.\n \n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=44982 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=44982 END:VEVENT BEGIN:VEVENT SUMMARY:CASTOR - Experiences\, status and plans DTSTART;VALUE=DATE-TIME:20081128T083000Z DTEND;VALUE=DATE-TIME:20081128T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-44794@cern.ch DESCRIPTION:\nDuring the last HEPiX meeting held at Taipei\, one of the pa rallel sessions was dedicated to CASTOR. Given the comprehensive view offe red to the audience it has been considered a good idea to rerun the presen tations\nat CERN. The presentations and structure will be the same\, that is:\n\nCASTOR status and plans\, by Sebastien PONCE\nSRM2 and monitoring p rojects in CASTOR\, by Dr. Giuseppe LO PRESTI\nCASTOR Operational Experien ces\, by Miguel COELHO DOS SANTOS\nIncreasing Tape Efficiency\, by Dr. Ste ven MURRAY\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=44794 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=44794 END:VEVENT BEGIN:VEVENT SUMMARY:How to suceed in implementing ITIL DTSTART;VALUE=DATE-TIME:20081217T093000Z DTEND;VALUE=DATE-TIME:20081217T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-46095@cern.ch DESCRIPTION:\nIT is a critical component of the scientific instruments the CERN is\nproviding for fundamental physics. Therefore\, the quality of se rvices\nprovided by the IT team is crucial and needs to be at its best. Bu t\,\nimproving the way IT is delivering services to the users is a\ndiffic ult task. Using ITIL is one way to improve the management of\nIT.\n\nBMC S oftware\, a leader in IT Management Software\, provides Business\nService Management solutions\, which ensures that IT consistently\nprovides the se rvices that deliver the greatest value for the\nenterprises. BMC Solutions help companies implement best practices\nsuch as ITIL.\n\nThe goal of thi s presentation is to show how BMC Solutions can help\nthe CERN improve the availability of IT services and improve the way\nIT deploys or update ser vices.\n\nAbout the speaker\n\nCyril Gobrecht is Business Solutions Manage r for Southern Europe at\nBMC Software. He has worked for 10 years as a so ftware consultant\nhelping customers improve the way they manage their IT. Cyril is ITIL\nService Manager and ITIL v3 Expert certified.\n\nhttp://in dico.cern.ch/conferenceDisplay.py?confId=46095 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=46095 END:VEVENT BEGIN:VEVENT SUMMARY:Scale-Out Storage and Beyond: Managing Diverse Data in an Modern E nvironment DTSTART;VALUE=DATE-TIME:20090306T140000Z DTEND;VALUE=DATE-TIME:20090306T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-50966@cern.ch DESCRIPTION:\nA reflection on the past\, present\, and discussion of futur e directions\nof data storage in this era of parallelized workflows. Infor mation will\nbe provided on NetApp's Scale-Out vision\, Data ONTAP GX\, ne w\ntechnologies such as solid state storage in an HPC environment\, and\nu pcoming advances in pNFS and federated file systems. Questions and\ncommen ts welcome.\n\nAbout the speaker\n\nRyan Sayre is a Consulting Systems Eng ineer for\nNetApp\,\na global\nleader in enterprise storage systems. He sp ecializes in Scale-Out\nStorage Architectures for HPC markets in Europe\, Middle East\, and\nAfrica. Ryan focuses on architecting and consulting on technical\napplications that utilize NetApp's clustered architectures\, Da ta\nONTAP GX and the upcoming Data ONTAP 8 Scale-Out platforms.\n\n\nBefor e joining NetApp\, Ryan spent four years as IT Infrastructure\nmanager at LAIKA Entertainment\, where he architected\, implemented\,\nand maintained a computing environment supporting multiple\nproductions in commercials\, video game development\, and feature\nfilms. He managed growth of the stu dio from 85 people to over 580\nutilizing over 300TB of storage and hundre ds of Linux and Windows\nrender nodes.\n\n\nRyan has also held positions i n the EDA industry\, SGI\, and Intel. He\nholds a bachelors degree in comp uter science from the\nUniversity of Portland\, and founded a 200 member I nternet\nCo-Operative.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?con fId=50966 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=50966 END:VEVENT BEGIN:VEVENT SUMMARY:Présentation de projets R&D HES-SO DTSTART;VALUE=DATE-TIME:20090320T130000Z DTEND;VALUE=DATE-TIME:20090320T163000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-43382@cern.ch DESCRIPTION:\nLe \nGITI\n(Groupement Professionnel des Ingénieurs en Tech nologies de l'Information) est une association professionnelle fondée le 20 septembre 2000 et compte à ce jour plus de 250 membres. L'ensemble de ses membres sont ingénieurs et actifs professionnellement dans un domaine des technologies de l'information.\n\n\nSon principal objectif est donc d e favoriser l'échange de bases de connaissances et d'expériences et cela sur trois axes:\n\nentre ses membres \nentre ses membres et les étudiant s des écoles d'ingénieurs \navec d'autres associations actives dans ses domaines d'activités\n\nNous souhaitons organiser une manifestation perme ttant de présenter une palette de projets de recherche appliquée actuell ement en cours dans les écoles du réseau HES-SO.\nCeci dans le but de fa voriser les la visibilité de ces écoles et offir aux équipes du CERN un e présentation sur les activités de Ra&D dans le domaine IT.\n\n\nSauf c ontre-indication\, les présentations sont données en français\n\n\n\nT élécharger le programme\n\n\nLes personnes externes au CERN sont priées de s'inscrire au moyen de ce formulaire\n\nhttp://indico.cern.ch/conferen ceDisplay.py?confId=43382 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=43382 END:VEVENT BEGIN:VEVENT SUMMARY:Datacenter Re-Evolution - Change Happens DTSTART;VALUE=DATE-TIME:20090327T133000Z DTEND;VALUE=DATE-TIME:20090327T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-55292@cern.ch DESCRIPTION:\n"Change is the law of life. And those who look only to the past or\npresent are certain to miss the future" - John F. Kennedy"\n\n\nT he Data Center landscape is changing at an incredible rate. Demand\nis inc reasing and technology is advancing rapidly\, more so than at\nany other t ime in our history.  Data Center operational cost\nincreases\, growing co nsumption\, and the corresponding carbon\nfootprint have increased the e xecutive visibility\, and pressure to\nget the environment under control. The world is now participating and\nthe demand on the datacenters to feed it will not stop. This\nRevolution requires an Evolution in thinking.  Th is session will dive\ninto the latest trends\, industry activities and tec hnologies that\ndatacenter owners and operators should be aware of.  Dean Nelson\, Sr\nDirector of Sun Microsystems Global Lab & Datacenter Design Services\,\nand co-founder of \nData Center Pulse\, an exclusive communit y of Data\nCenter owners and operators\, will be sharing information on Su n's\nlatest datacenter innovations as well as updates on the Data Center\n Pulse \nactivities\, \nincluding \n\nThe Chill Off\,\nFocus Tracks\, an d more.\n\nAbout the speaker\n\nDean is the Sr. Director of Global Lab & D atacenter Design Services (GDS) in the Work Environments (WE) business uni t of Sun Microsystems. GDS bridges the gap between Facilities & IT/Enginee ring and is responsible for managing Sun's multi-billion dollar global tec hnical infrastructure portfolio including datacenter design\, standards an d strategy. The GDS work resides in the Act portion of Sun's Eco strategy\ , and was showcased at Sun's Eco Launch in August\, 2007\, and the Colorad o Data Center grand opening in January\, 2009. Because of overwhelming cus tomer interest\, the GDS processes and design became a Sun service offerin g through the Data Center Efficiency practice (DCE) in December\, 2007.\n\ n\nDean has been in the technology industry for 19 years\, of which 16 hav e been with Sun. He spent four years in Sun manufacturing in roles ranging from component level debug to managing quality. This included helping the drive to achieve ISO 9002 certification. Dean joined the Sun Engineering community in 1993. He led systems and network administration support for s ome of Sun's largest and most complex R&D lab environments.\n\n\nDean left Sun in 2000 to join a networking startup company called Allegro Networks. At Allegro\, he built a world-class QA team\, state of the art global R&D lab environments and fully integrated automation system.\n\n\nIn 2003\, D ean returned to Sun joining the newly formed N1 organization. He orchestra ted the integration of Terraspring and Center Run R&D labs and the merger of Sun Cluster into N1. In mid 2003\, Dean took over management all of N1' s R&D labs\, build engineering\, automation\, and capital budget responsib ilities world-wide. In 2004\, he became a leading member of the Global Lab & Datacenter Design Services (GDS) team tasked with creating a strategy t o standardize Sun's multi-billion dollar technical infrastructure portfoli o. He was the architect of the GDS operating model and lead design enginee r for lab and datacenter projects world-wide. In 2006\, he became the GDS Director. Since taking over the group\, Dean has delivered the GDS strateg y\, executing the largest technical infrastructure consolidation in Sun's history compressing the 1.4 million square foot portfolio by 48%\, a $250M dollar investment.\n\n\nDean is the Founder of Data Center Pulse - an exc lusive datacenter owner community influencing the industry through the len s of the end user. He is an author of the Sun Blueprint\, "Energy Efficien t Datacenters: The Role of Modularity in Datacenter Design" and was featur ed in Contrarian Minds\, a focus on the engineers\, scientists and dreamer s of Sun Microsystems. Dean is also the recipient of Sun's prestigious Inn ovation Award presented by Sun's CEO\, Jonathan Schwartz and CTO\, Greg Pa padopoulos in July\, 2008.\n\n\nDean lives in Northern California with his wife and daughter.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId =55292 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=55292 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Prague CHEP'09 Conference DTSTART;VALUE=DATE-TIME:20090424T073000Z DTEND;VALUE=DATE-TIME:20090424T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-53985@cern.ch DESCRIPTION:\nThis year\, the 17th\n\nComputing in High Energy and Nuclear Physics\n(CHEP) conference has been held at Prague\, Czech Republic\, fro m 21-27 March 2009.\n\n\nCHEP conferences provide an international forum t o exchange information on computing experience and needs for the High Ener gy Physics and Nuclear Physics communities\, and to review recent\, ongoin g and future activities. \n\n\nThe speakers at this Computing Seminar will present their views of the CHEP'09 conference.\n\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=53985 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=53985 END:VEVENT BEGIN:VEVENT SUMMARY:LinuxDays 2009 "openforbusiness" DTSTART;VALUE=DATE-TIME:20090519T120000Z DTEND;VALUE=DATE-TIME:20090519T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-58974@cern.ch DESCRIPTION:\nLinuxDays is the first IT show in French-speaking Switzerlan d to be\ndedicated to Open Source solutions. \nLinuxDays 2009\nis the 5th edition in the series\, and will take place in Geneva on June 3 - 5.\n\n\n Born 25 years ago\, Open Source Software is no longer a geek hobby\nonly\; on the contrary it has become a development model which has\nbeen proved and which is particularly attractive\, especially with the\neconomic crisi s. All the big software vendors are producing and\nworking with Open Sourc e at some point\, as you can see it at\nMicrosoft or Oracle. LinuxDays are aiming at promoting Open Source\nsolutions to businesses and organisation s\, by introducing the\nexisting solutions\, by animating debates\, by gat hering all the\ncompanies that have the best competencies for your Open So urce\nprojects. The event is open to everyone.\n\n\nThis year\, LinuxDays are organised collaboratively with the State of\nGeneva\, Canonical (vendo r of the popular Ubuntu Linux distribution)\,\nErnst & Young\, the Linux P rofessionnal Institute (LPI) and the\nEnabling Grids for E-sciencE project (EGEE). We are expecting more\nthan 1000 visitors and 40 conferences. Our key note speaker\, Michel\nRocard\, ex Prime Minister of France and a str ong adversary to\nsoftware patents\, will give a conference on Friday\, Ju ne 5th at\n14:00. There will also be workshops to continue the talks and i deas\nstarted in the conferences\, and LPI certifications will be offered\ nwith a discount for all visitors.\n\n\nThis presentation will focus on th e highlights of the planned event\, offering attendants a glimpse of what it is to be expected.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?conf Id=58974 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=58974 END:VEVENT BEGIN:VEVENT SUMMARY:Efficient 40 Gb/s Routing for CERN’s Networks DTSTART;VALUE=DATE-TIME:20090616T080000Z DTEND;VALUE=DATE-TIME:20090616T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-60280@cern.ch DESCRIPTION:\nHyperchip's H40G Green NGI router provides the densest and m ost energy-efficient IP/MPLS routing\, is roughly 3x more power-efficient and an order of magnitude denser than traditional routers. The H40G provi des standard routing protocols\, and the control plane is extensible to ru n experimental routing protocols on their own processors. The H40G is sca lable from the initial 1U MicroTCA systems to ATCA and multi-chassis syste ms.\nThe heart of the H40G is a leading-edge network processor based on re configurable computing - using an FPGA to deliver full-featured line rate network processing more efficiently than software-programmable network pro cessors.\n\n\nSince CERN processes\, as well as distributes massive amount s of data\, this raises the possibility of using extra FPGA resources for custom processing when the data is passing through the H40G\, or of H40G h ardware with dedicated physics data processing pipelines. Hyperchip has pi oneered pipeline architectures that would likely be applicable to such pro cessing.\n\n\nThe discussion will thus cover the basics of the H40G\, CERN 's layer-3+ routing requirements\, and the possibility of custom CERN proc essing on H40G systems.\n\nAbout the speakers\nRichard Norman\, co-founder and CTO of Hyperchip\n\nRichard Norman is co-founder and chief technology officer of 4198638 Canada Inc. (Hyperchip). Richard lead Hyperchip from i ts roots in chip architecture to extreme performance network systems manuf acturer\, he is committed to bringing end-to-end gigabit connectivity to n etwork users everywhere.\n\n\nRichard has devoted nearly 10 years of paten t work on semiconductor innovations\, massively parallel semiconductors an d extreme performance computing – all towards building the world’s fir st router scalable to petabit levels. Currently\, he has 11 patents issued and 27 pending.\n\n\nBefore founding Hyperchip\, Richard co-founded Profi table Entertainment Products\, where he was president and director of theo retical research\, specifically responsible for prototype and product deve lopment. Prior to this\, he was senior programmer at Vermont Creative Soft ware\, Inc.\n\n\nUnderlying these professional achievements is 35 years of programming experience\, including eight years at IBM as systems programm er\, where he received the Technology Division’s highest award for innov ation. To complement his software expertise\, he also has more than 10 yea rs of hardware system design experience\, including the design of massivel y parallel systems\, fault–tolerant architectures and integrated systems .\n\n\nRichard is a dynamic and accomplished speaker who has presented or chaired sessions at over 30 conferences\, on subjects ranging from network migration strategies to supercomputing architectures. He holds a bachelor of science degree in physics from Clarkson University in New York.\n\nMir ek B. Chochol\, responsible at Hyperchip for International Business Develo pment\n\nMirek B. Chochol is responsible for international marketing and b usiness development related to the new H40G Router and the existing PBR-12 80 IP Core System. The Hyperchip H40G Green NGI Router is a 40Gb/s router over 10 times smaller and a fraction of the cost of competitors\, using in dustry standard MicroTCA chassis and open-standard interfaces.\n\n\nMirek has 20 years experience in the telecommunications industry. \n\n\nBefore j oining Hyperchip\, Mirek was Executive Director and Head of International Business\, Marketing and Sales Development in EMEA and Asia at Nortel Netw orks and VP of International / National Business & Sales Development for C anada and EMEA region at Realm Networks.\n\n\nA key achievement by Mirek w as annualized sales revenues from two (2) million dollars (US) to over one hundred and fifty (150) million dollars (US).\n\n\nMirek is a multilingua l professional with broad international business experience in Start-ups\, SMEs and Blue-chips. He is a perceptive\, disciplined and dedicated manag er poised to tackle complex challenges in the rapidly changing field of na tional and international Telco networks infrastructure. Having worked with some of the best minds in the global telecoms industry\, and with many le aders in global service provision\, his expertise can kick start new techn ology initiatives\; drive new product or service revenue streams\; increas e margin\, market penetration or sales revenue acceleration.\n\n\nMirek st arted his professional career as senior engineer / scientist at Northern T elecom Corporation and in BNR\, and was responsible for the design and dev elopment of world first digital video/audio codec DV45\, as well as to wri te technical and test specifications and the development of military commu nications equipment.\n\n\nMirek holds a Masters of Engineering degree from CVUT (Czech University of Technology)\, and is a member of the SMPTE. Mir ek Holds dual citizenship\, Canadian and European. \n\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=60280 LOCATION:CERN 513-1-024 URL:http://indico.cern.ch/conferenceDisplay.py?confId=60280 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Spring 2009 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20090619T120000Z DTEND;VALUE=DATE-TIME:20090619T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-60279@cern.ch DESCRIPTION:\nThe Spring 2009 meeting of HEPiX was held at Umea University \, Sweden on 25-29 May 2009. HEPiX is a global organization of service man agers and support staff providing computing facilities for the High Energy and Nuclear Physics community. \n\n\nThe speakers will present a summary of the subjects addressed during the meeting:\nGeneral highlights - Helge Meinhard\nMichal Kwiatek\nArne Wiebalck\n\n\nhttp://indico.cern.ch/confere nceDisplay.py?confId=60279 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=60279 END:VEVENT BEGIN:VEVENT SUMMARY:Evolution of the Windows Kernel Architecture - Win7 and Beyond DTSTART;VALUE=DATE-TIME:20090702T120000Z DTEND;VALUE=DATE-TIME:20090702T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-62731@cern.ch DESCRIPTION:\nThe core architecture of the Windows kernel differs signific antly from UNIX\,\nthough many of the functions are necessarily similar.  This talk will cover the\nWindows kernel architecture\, focusing on areas of difference from UNIX.  Then\nsome specific recent kernel changes will be described: integrity levels\, removalof the hottest kernel locks\, and the move towards user-mode scheduling to\nsupport the ConcRT task-based c oncurrency run-time.  Finally\, the talk will lookat some of the OS chall enges ahead due to the silicon power-wall and the\nresulting heterogeneous multi-core/many-core systems that will dominate the PC\nlandscape within a few years.\n\nAbout the speaker\n\nDave Probert is a kernel architect wi thin the Windows Core Operating Systems\nDivision at Microsoft where he is currently working on the next generations of\nWindows. He is also the arc hitect for the Windows Academic Program\, developing\nboth the WRK package and ProjectOZ. Previously he managed kernel development for Windows\, sta rting with the Windows 2000 release. Dave Probert joined Microsoft\nin 199 6\, after earning his Ph.D. in Electrical & Computer Engineering at UC\nSa nta Barbara developing the SPACE project with Prof. John Bruno. His prior\ nindustry experience includes serving as Vice President of Software Engine ering\nat Culler Scientific Systems\, consulting for various companies on UNIX kernel\ninternals\, and working as a systems architect at Burroughs c orporation designinghardware and writing microcode for the B1900. During S pring Quarter 2009 Dave\nProbert taught the undergraduate operating system s course at University of\nWashington's Bothell campus using Windows.\n\n\ nhttp://indico.cern.ch/conferenceDisplay.py?confId=62731 LOCATION:CERN Main Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=62731 END:VEVENT BEGIN:VEVENT SUMMARY:C++: New and Improved! DTSTART;VALUE=DATE-TIME:20090722T123000Z DTEND;VALUE=DATE-TIME:20090722T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-55999@cern.ch DESCRIPTION:\nThis talk summarizes the additions and changes forthcoming i n C++0X\, the next C++ standard. After briefly reviewing the C++\nstandard s committee's approach\, goals\, and anticipated timeline for C++0X\, we h ighlight several dozen language and library features\,\nincluding concepts \, concurrency\, rvalue references\, and uniform\ninitialization syntax.\n \nAbout the speaker\n\nWith degrees in both mathematics and computer scien ce\, Dr. Brown has\nbeen a professional programmer for nearly 40 years.\n\ n\nHis experience spans both academia and industry: In academia\, he\nfou nded a Computer Science Department\, served as Professor and\nDepartment H ead\, and taught and mentored at all levels. In industry\, he\nmanaged an d mentored programming staff for a computer reseller\, and was self-employ ed as a software consultant and commercial trainer.\n\n\nDr. Brown has wor ked for Fermilab since 1996. He is now part of the\nComputing Division's Future Programs and Experiments Quadrant\,\nspecializing in C++ consulting and programming. He participates in the\ninternational C++ standardizati on process and is responsible for several\naspects of the forthcoming upda ted C++ Standard. In addition\, he is the\nProject Editor for the forthco ming C++ Standard on Mathematical Special\nFunctions.\n\n\nAbout himself\, Dr. Brown writes: "Be forewarned: Based on my training\nand my experienc e\, I hold some rather strong opinions about computer\nsoftware and progra mming methodology. While these opinions are not\nshared by all programmer s\, they should be! :)"\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?co nfId=55999 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=55999 END:VEVENT BEGIN:VEVENT SUMMARY:Securing with the OSSTMM 3 DTSTART;VALUE=DATE-TIME:20090904T080000Z DTEND;VALUE=DATE-TIME:20090904T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-62747@cern.ch DESCRIPTION:\nThe \nOpen Source Security Testing Methodology Manual is wel l on its way\nof being part of the ISO standard for security and a world-r ecognized\nstandard in its own right. However many of its concepts are not only\nnew but contradict accepted best practices and the current security \nbody of knowledge. This seminar will show you how this came to be\, why\ nit's correct\, and how the OSSTMM will allow you to not only find\nperfec t security but also to analyze and measure it.\n\n\nIf there is interest\, a second part can be held giving more practical examples\, go into more d epth\,\nand conduct more Q&A. If you are interested please contact Andreas Unterkircher.\n\nAbout the speakers\n\nPete Herzog is the creator of the OSSTMM and lead researcher behind\nthe new change from solution-based secu rity testing to operational\nsecurity testing\, security metrics\, and tru st metrics.\n\n\n\nNicolas Mayencourt is the CEO of Dreamlab\nTechnologies AG\, affiliate partner of ISECOM in Switzerland.\n\n\nhttp://indico.cern. ch/conferenceDisplay.py?confId=62747 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=62747 END:VEVENT BEGIN:VEVENT SUMMARY:An Introduction to Drupal 7 DTSTART;VALUE=DATE-TIME:20090907T083000Z DTEND;VALUE=DATE-TIME:20090907T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-66161@cern.ch DESCRIPTION:What is Drupal (and what isn't it)?\nHow is Drupal maintained? \nWhy is Drupal amazing?\nWhere is Drupal going?\nAngela Byron is the Drup al 7 core maintainer\, recipient of the Google-O'Reilly Open 2008 Source A ward for Best Contributor\, and an Open Source evangelist who lives and br eathes Drupal. She got her start as a Google Summer of Code student in 200 5 and since then has completely immersed herself in the Drupal community. Her work includes core coding and patch review\, creating and contributing modules and themes\, testing and quality assurance efforts within the pro ject\, improving documentation\, and providing user support on forums and IRC.\n\nAngela has lead Lullabot's collaboration on O'Reilly's first Drupa l book\, entitled "Using Drupal". She is on the Board of Directors for the Drupal Association and helps drive community growth by leading initiative s to help get new contributors involved such as Drupal’s participation i n Google Summer of Code and Google’s Highly Open Participation (GHOP) pr ograms. She is a sought-after lecturer on many themes\, especially the top ics of women in Open Source and just about anything to do with Drupal. Ang ie is known as "webchick" on drupal.org.\n\nNote: the location of this sem inar has not yet been confirmed. Please email the organiser to express an interest and be kept up to date\, or apply to attend this event.\n\n\nLink s\nDrupal.org\nVideo: What is Drupal?... in 57 seconds\nUsing Drupal on Am azon.com\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=66161 LOCATION:CERN Room B URL:http://indico.cern.ch/conferenceDisplay.py?confId=66161 END:VEVENT BEGIN:VEVENT SUMMARY:Social networks - From virtual to real life DTSTART;VALUE=DATE-TIME:20090930T090000Z DTEND;VALUE=DATE-TIME:20090930T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-67964@cern.ch DESCRIPTION:\nToday major websites and social networks such as facebook an d myspace were initially designed to be used by a home-alone user communic ating from a desktop PC. This leads to what we call a virtual social life where people enjoy being connected with their friends in order to share pi ctures\, videos and messages while at home (office?!?).\nOn the other hand \, the iPhone and other smartPhones create new ways to use the Internet wh ile on the road\, where ubiquity and positioning(GPS) become two\nessentia l ingredients that were not taken into account by first generation\nsocial networks.\n\n\nAt MadeinLocal.com\, \na Swiss start-up from EPFL\, we bel ieve that social networks\nshould help the user to have a great social lif e.\nTherefore\, we specialized our social network to provide useful inform ation for\nthe user's real social life\, his local life. In particular\, w e put together\nlocal information about the most recommended must-see attr actions\, bars\,\nrestaurants and events and we bring him tools to invite and communicate with hisfriends in order to go out and enjoy with friends wonderful real life moments.\n\nAbout the speaker\n\nManuel Acevedo is the CEO of MadeinLocal.com\, an EPFL start-up specialized in new generation s ocial networks.\nAfter his IT studies in Switzerland\, France and the US\, he worked for 6 years on high security constraint projects for the Canton of Geneva pubic administration. In particular the e-voting system and maj or Police projects.\n \nIn 2008\, he started MadeinLocal.com in a "garage" in Canton de Vaud. Initial prototypes illustrating the team initial visio ns on the evolution of social networks led to several\nlocal and national prizes on innovation (EPFL innogrant\, Venture Kick I & II).\nMadeinLocal goes public these days in Lausanne where University students are\ncurrentl y participating in the Best Lausanne Bar Election (by the EPFL-UNIL\nsocia l network).\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=67964 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=67964 END:VEVENT BEGIN:VEVENT SUMMARY:Massive Data Computing in a Connected World DTSTART;VALUE=DATE-TIME:20091005T090000Z DTEND;VALUE=DATE-TIME:20091005T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-68743@cern.ch DESCRIPTION:\nComputing platforms and the World Wide Web are undergoing si gnificant architectural transitions driven by the unprecedented convergenc e of the need to process massive amounts of data with the availability of massive compute resources. This has significant algorithmic implications f or traditional approaches to many common computational problems in visual computing and analytics. This further has the potential to enable a new cl ass of Connected Computing applications. The service-oriented focus and re al-time nature of these emerging applications make computing more implicit and capable of delivering a significantly more immersive experience to a broader class of end-users. This talk will explore the compute-platform im plications of this trend.\n\nAbout the speaker\n\nPradeep Dubey is a Senio r Principal Engineer and Director of Throughput Computing Lab\, part of Mi croprocessor and Programming Research division of Intel Labs. His research focus is computer architectures to efficiently handle new application par adigms for the future computing environment. He is currently responsible f or throughput computing applications on Intel’s Larrabee architecture. D ubey previously worked at IBM's T.J. Watson Research Center\, and Broadcom Corporation. He was one of the principal architects of the AltiVec* multi media extension to Power PC* architecture. He also worked on the design\, architecture\, and performance issues of various microprocessors\, includi ng Intel® i386TM\, i486TM\, and Pentium® processors. He holds 26 patents and has published extensively. Dr. Dubey received his PhD in electrical e ngineering from Purdue University. He is a Fellow of IEEE.\n\n\nhttp://ind ico.cern.ch/conferenceDisplay.py?confId=68743 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=68743 END:VEVENT BEGIN:VEVENT SUMMARY:Getting it all with C++: Abstraction\, Reusability\, Performance a nd Future-Safety DTSTART;VALUE=DATE-TIME:20091007T140000Z DTEND;VALUE=DATE-TIME:20091007T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-59152@cern.ch DESCRIPTION:\nThe talk will introduce techniques based on current and futu re C++\nfeatures to enable taking advantage of a processors full potential and\npotentially even the use of accelerators and other evolving technolo gies\nwithout compromising the usability of the library interfaces.\n\n\nh ttp://indico.cern.ch/conferenceDisplay.py?confId=59152 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=59152 END:VEVENT BEGIN:VEVENT SUMMARY:Scalla/xrootd 2009 Developments DTSTART;VALUE=DATE-TIME:20091012T140000Z DTEND;VALUE=DATE-TIME:20091012T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-66432@cern.ch DESCRIPTION:\nThis talk will describe recent additions to the\nScalla/xroo td\nsoftware suite: File Residency Manager (FRM)\, torrent transfers\, aut omatic\nmonitor reporting\, ephemeral files\, and Simple Servert Inventory SSI).\nAdditionally\, the talk will discuss issues with Solid State Devic e support\nand future directions in that area.\n\n\nhttp://indico.cern.ch/ conferenceDisplay.py?confId=66432 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=66432 END:VEVENT BEGIN:VEVENT SUMMARY:Science data processing in the Cloud: first experience in Gaia DTSTART;VALUE=DATE-TIME:20091028T133000Z DTEND;VALUE=DATE-TIME:20091028T143000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-70442@cern.ch DESCRIPTION:\nWith the maturing of Cloud Computing\, it is now feasible to run scientific\napplications in the cloud. Data storage and high performa nce computing resourcesare fundamental for scientific applications. Outsou rcing these services\nleverages scalability\, flexibility\, high availabil ity at lower prices compared\nwith traditional in-house data processing.\n \n\nThe Server Labs\nhas developed and driven a highly successful feasibil ity study\non behalf of the European Space Agency which identified clear b enefits of\nrelocating\nGaia's\ndistributed astrometric processing  to th e Cloud. This session evaluates Amazon's EC2/S3 suitability for this scena rio. The aim is to demonstrate how cloud computing systems can be a viable and cost-effective\nsolution for HPC applications. A new study is current ly underway  for Gaia to test a new simulated dataset with 60 million pri mary stars and 1TB of data\, and scale out to 1000 Amazon EC2 nodes.\n\nAb out the speaker\n\nPaul Parsons is Co-Founder & CTO of The Server Labs. He has over 20 years of\nexperience in software development and was senior s oftware engineer in the\nEuropean Space Agency (ESA). After ESA\, he worke d in the Deutsche Boerse\,\nCommerzbank and DG Bank\, designing and implem enting real time distributed\nsystems. Since then he has been an architect in different industries\, designing\nhigh performance solutions. In 2004 he co-founded The Server Labs and has been\nactively involved technically in many clients\, including the Gaia project for\nESA.\n\n\nhttp://indico. cern.ch/conferenceDisplay.py?confId=70442 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=70442 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Autumn 2009 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20091120T083000Z DTEND;VALUE=DATE-TIME:20091120T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-71771@cern.ch DESCRIPTION:\nThe Autumn 2009 meeting of HEPiX will be held at LBNL\, Berk eley USA on 26-30 October 2009. HEPiX is a global organization of service managers and support staff providing computing facilities for the High Ene rgy and Nuclear Physics community. \n\n\nThe speakers will present a summa ry of the subjects addressed during the meeting:\nGeneral highlights - Hel ge Meinhard\nJuraj Sucik\nUlrich Schwickerath\n\n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=71771 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=71771 END:VEVENT BEGIN:VEVENT SUMMARY:The Data Warehouse Challenge at Credit Suisse - Prepared for the N ext Lehman Case? DTSTART;VALUE=DATE-TIME:20091130T100000Z DTEND;VALUE=DATE-TIME:20091130T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-72339@cern.ch DESCRIPTION:\nOne of the most essential items of a bank is data - it shoul d be stored securely and should never be lost. This data contains informat ion about customers (private and corporate)\, banking products (such as op tions\, futures and derivatives)\, stock information\, trading information as well as market research data. The amount of data that is stored in Cre dit Suisse's enterprise data warehouse is close to 0.5 petabytes.\n\n\nIn this talk the speaker will explain two different strategies of how Credit Suisse handles these challenges\, namely via a high availability architect ure and via semantic data warehouse search. The first half of the talk sho ws results from a detailed proof of concept based on the new features of O racle 11g Active Data Guard to replicate data between two distributed data centers along Lake Zurich. The second half of the talk focuses on a joint research project with ETH Zurich where we combine technologies from three areas (databases\, search and semantic web) to tackle the Lehman Case\, i .e. to quickly identify all bank products related to the crashed Lehman in vestment bank. The main goal of this research project is to enable intuiti ve\, self-service business intelligence (BI) for business analysts to expl ore a multi-terabyte enterprise warehouse without requiring in-depth IT kn owledge.\n\nAbout the speaker\n\nDr. Kurt Stockinger is a data warehouse ( DWH) architect at Credit Suisse\, Zurich since fall 2007. He currently wor ks on designing and prototyping DWH algorithms for a terabyte-scale enterp rise warehouse with focus on (bi-) temporal database techniques. Prior to joining Credit Suisse\, Kurt worked for 4 years at Berkeley Lab\, Universi ty of California performing research on multi-dimensional indexing and que ry methods for large-scale scientific data as well as high-performance vis ual analytics (query-driven visualization on modern supercomputers). From 2000 to 2002 Kurt was heading the Replica Optimization Team of the EU Data Grid Project at CERN. In 2008 he received an R&D 100 Technology Award tog ether with three colleagues from Berkeley Lab for his research on FastBit – a multi-dimensional bitmap index engine that was applied for query acc eleration in the areas of high-energy physics\, astrophysics\, and combust ion simulation as well as computer network traffic analysis.\n\n\nhttp://i ndico.cern.ch/conferenceDisplay.py?confId=72339 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=72339 END:VEVENT BEGIN:VEVENT SUMMARY:The Challenges of Exascale Computing DTSTART;VALUE=DATE-TIME:20091204T130000Z DTEND;VALUE=DATE-TIME:20091204T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-75209@cern.ch DESCRIPTION:http://indico.cern.ch/conferenceDisplay.py?confId=75209 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=75209 END:VEVENT BEGIN:VEVENT SUMMARY:Vulnerability Assessment and Secure Coding Practices for Middlewar e DTSTART;VALUE=DATE-TIME:20091207T090000Z DTEND;VALUE=DATE-TIME:20091207T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-74803@cern.ch DESCRIPTION:\nSecurity is crucial in the software that we develop and use. This\ntutorial is relevant to anyone wanting to learn about assessing sof tware\nfor security flaws and for developers wishing to minimize security flaws\nin software they develop.\nWe share our experience in vulnerability assessment of grid middleware.\nYou will learn skills critical for develo pers and analysts concerned\nabout software security\, and the importance of independent vulnerability\nassessment.\n\n\nThe first part of this tuto rial covers a process to actively discover\nvulnerabilities. We show how t o gather information about a system and\nhow to use this to direct the sea rch for vulnerabilities\, and how to\nintegrate this into the development cycle.\n\n\nThe second part of this tutorial examines coding practices to prevent\nvulnerabilities by describing more than many types of vulnerabili ties\nwith examples of how they commonly arise\, and techniques to prevent \nthem. Most examples are in C and C++.\n\nAbout the speakersBarton Miller \n\nBarton Miller is Professor of Computer Sciences at the University of\n Wisconsin\,\nMadison. He directs the Paradyn Tool project\, which is inve stigating binary\ncode instrumentation and analysis technologies for high performance\ncomputing\nand cyber-security applications. He also directs the MIST vulnerability\nassessment project in collaboration with the Auton omous University of\nBarcelona.\nMiller has published widely in the top co mputer security and high\nperformance\ncomputing conferences. He founded t he field of fuzz testing\, widely used\nin the software engineering and co mputer security fields.\n\n\nMiller co-chaired the Supercomputing 2008 Tec hnical Program Tutorials\,\nand is co-chair of the upcoming 2010 Dagstuhl Seminar on Program Development\nfor Extreme-Scale Computing. Miller has b een on the editorial boards of\nIEEE\nTransactions on Parallel and Distrib uted Systems\, the International\nJournal of\nParallel Processing\, Concur rency and Computation Practice and\nExperience\, and\nComputing Systems.\n \n\nMiller is the chair of the IDA Center for Computing Sciences Program R eview\nCommittee\, has been on the Los Alamos National Laboratory Computin g\nCommunications and Networking Division Review Committee\, U.S. Secret\n Service Electronic Crimes Task Force (Chicago Area)\, and the Advisory\nBo ard for the International Summer Institute on Parallel Computer\nArchitect ures\, Languages\, and Algorithms in Prague.\n\n\nMiller received his Ph.D . degree in Computer Science from the University of\nCalifornia\, Berkeley in 1984. He is a Fellow of the ACM.\n\nElisa Heymann\n\nElisa Heymann is Associate Professor of operating systems at the\nUniversidad Autonoma of Barcelona (Spain). She attained her B.S. degree in computer science in 199 2 at the University Simon Bolivar (Venezuela)\, and her MSc and PhD degree s from the University Autonoma of\nBarcelona in 1995 and 2001\, respective ly. Her research interests are in\nthe area of resource management\non dis tributed systems\, middleware for distributed systems/grid systems\nand co mputer security. She has participated in several research projects related with the development of management strategies for\nparallel applications on distributed environments and vulnerability\nassessment for middleware.\ n\n\nShe is author or co-author of a significant number of papers in\nscie ntific journals and symposiums.\n\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=74803 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=74803 END:VEVENT BEGIN:VEVENT SUMMARY:Shibboleth DTSTART;VALUE=DATE-TIME:20091214T100000Z DTEND;VALUE=DATE-TIME:20091214T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-76599@cern.ch DESCRIPTION:This talk will give a basic introduction to the components of \nShibboleth\nand the \nSAML\nprotocols that it implements. After which t he primary focus will be on the component used to protect resources (as op posed to the component which performs the actual authentication and is int egrated with an organization's identity management system). Finally\, the talk will wrap up with some troubleshooting suggestions.\n\nAbout the spe aker\n\nChad La Joie is a software engineer within the Grid and AAI teams at\nSWITCH\, Switzerland's national research and education network\norgani zation. In addition to his position at SWITCH he also serves as\nthe techn ical co-lead of the Shibboleth project. His primary work\nefforts focus on identity management systems and improving the security\nof distributed sy stems. In addition he has worked in the areas of\nbio-informatics\, natura l language processing\, disease and biological\nagent transmission detecti on and modeling\, and various other topics.\n\n\nhttp://indico.cern.ch/con ferenceDisplay.py?confId=76599 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=76599 END:VEVENT BEGIN:VEVENT SUMMARY:A Fresh Perspective on Distributed Scientific Applications and Cyb erInfrastructure DTSTART;VALUE=DATE-TIME:20100108T130000Z DTEND;VALUE=DATE-TIME:20100108T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-76961@cern.ch DESCRIPTION:\nIt is generally accepted that the ability to develop large-s cale distributed\napplications that are extensible and independent of infr astructure details has\nlagged seriously behind other developments in cybe rinfrastructure. As the\nsophistication and scale of distributed infrastru cture increases\, the complexityof successfully developing and deploying d istributed applications increases bothquantitatively and in qualitatively newer ways.\n\n\nIn this talk\, we will present a\nfresh perspective to di stributed applications and distributed\ncyberinfrastructure using SAGA. SA GA provides the basic and most commonly used\nfunctionality for a broad ra nge of distributed applications and provides the\nbuilding blocks upon whi ch to construct higher-levels of functionality and\nabstractions. We will show how a SAGA-based approach can address many common\nbarriers in the de velopment and execution of distributed applications\, while\nfacilitating the development of applications that by design are --\nInteroperable\, Sca le-Out\, Extensible\, Adaptive and Simple.\n\nAbout the speaker\n\nShanten u is the Director for Cyberinfrastructure Development at the\nCenter for C omputation and Technology (CCT)\,\nand a Research Professor in Computer Sc ience at Louisiana State University\n(LSU). He is also a theme-leader at t he e-Science Institute\, Edinburgh and\na Visiting Researcher at UCL. His research interests lie at the triple\npoint of Computational Science\, Cyb erinfrastructure Development and\nComputer Science.  His Research is supp orted by a broad range of sources\,\nincluding UK-EPSRC\, US-NSF and NIH a s well with SUN and Google (GSOC). He\nis a co-PI for $2.6M NSF award that enables LSU/LONI participation as a\nTeraGrid Resource Provider. Shantenu leads the SAGA project and is\ncurrently working on writing a book on "Ab stractions for Distributed\nApplications and Systems: A Computational Scie nce Perspective" to be\npublished by Wiley in 2010.\n\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=76961 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=76961 END:VEVENT BEGIN:VEVENT SUMMARY:Code Analysis Tools: finding your bugs before somebody else does! DTSTART;VALUE=DATE-TIME:20100219T083000Z DTEND;VALUE=DATE-TIME:20100219T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-80138@cern.ch DESCRIPTION:\nSoftware security is crucial for keeping the organization se cure. One\nway of making software developed at CERN more secure is using s ource\ncode analysis tools. Such tools statically look at source code\, an d\nproduce a list of potential bugs found\, including potential security\n flaws. Running these tools requires minimal effort\, and while obviously\n they cannot find all existing issues\, they are helpful at finding the\nmo st common ones - so the cost/benefit ratio is very good.\nThis presentatio n\, aimed mostly at CERN developers\, will discuss a\ncouple of tools sugg ested for C/C++\, Java\, PHP\, Perl and Python. It will\ninclude a demonst ration of how these tools can be easily used.\n\n\nhttp://indico.cern.ch/c onferenceDisplay.py?confId=80138 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=80138 END:VEVENT BEGIN:VEVENT SUMMARY:StarBED: A Large-scale Network Experiment Testbed DTSTART;VALUE=DATE-TIME:20100226T100000Z DTEND;VALUE=DATE-TIME:20100226T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-83203@cern.ch DESCRIPTION:\nThe \n\nNational Institute of Information and Communications Technology (NICT)\nin Japan manages a large-scale network experiment test bed called StarBED\,\nthat is located at the Hokuriku Research Center\, in Ishikawa prefecture.\nStarBED counts more than 1000 PCs that have been us ed over the years for\nvarious network experiments and research. We shall present the\ninfrastructure and experiment-support software of StarBED\, a s well as some\nof the research projects that were carried out on StarBED. We shall also\npropose several scenarios that show how StarBED could be u seful for\nCERN-related activities\, such as testing LHC software in a con trolled\nenvironment before deployment in the production network.\n\nAbout the speaker\n\nRazvan Beuran is researcher with the National Institute of Information\nand Communications Technology\, Hokuriku Research Center\, I shikawa\,\nJapan. In 2006 he was post-doc research fellow with the Japan\n Institute of Science and Technology\, Ishikawa\, Japan\, where he is\ncurr ently project researcher. From 2001 to 2005 he was with CERN\,\nGeneva\, S witzerland as researcher\, and then as project associate. From\n1999 to 20 05 he was also with "Politehnica" University\, Bucharest\,\nRomania as res earch assistant\, and then as teaching assistant.\n\n\nHis research topics include: quality testing and measurement in wired\nand wireless networks\ , network emulation\, and network reliability and\ndependability in connec tion with disaster situations. He received the\njoint Ph.D. degree in Elec trical Engineering and Computer Science from\n"Politehnica" University\, B ucharest\, Romania and "Jean Monnet"\nUniversity\, Saint Etienne\, France in 2004.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=83203 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=83203 END:VEVENT BEGIN:VEVENT SUMMARY:OpenCL\, Design Patterns and Software for heterogeneous many core platforms DTSTART;VALUE=DATE-TIME:20100315T130000Z DTEND;VALUE=DATE-TIME:20100315T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-87133@cern.ch DESCRIPTION:Future computer systems will build on a many core foundation. The quest to optimize performance per watt will push these cores to specia lize\, so these will be heterogeneous many core systems. The hardware tren ds are clear\; it's the software response that is a bit murky. Software mu st change to take advantage of these new platforms. This is a technical pr oblem\, but it's also a social problem as the software must address the ne eds of different application and programmer communities.\n\nStandards such as OpenCL are a critical part of the solution. This new programming langu age lets programmers write software that runs on CPUs\, GPUs or combinatio ns thereof. A long term solution\, however\, will only emerge from a clear understanding of how programmers architect parallel software. This is the focus of my research on Design Patterns carried out in collaboration with the ParLab at UC Berkeley. In this talk\, I will describe these two proje cts and how they work together to address many core programming challenges .\nAbout the speaker\nTim Mattson is an applications programmer (Ph.D. Che mistry\, UCSC\, 1985). He does linear algebra\, finds oil\, shakes molecul es\, solves differential equations\, and models electrons in simple atomic systems. He has spent his career working with computer scientists to make sure the needs of parallel applications programmers are met.\nDr. Mattson is also engaged in a long term research program with UC Berkeley's ParLab on abstractions that bridge across parallel system design\, parallel prog ramming environments\, and application software. This work builds on his b ook "Design Patterns in Parallel Programming" (written with Professors Bev erly Sanders and Berna Massingill). The patterns provide the "human angle" and help keep his research focused on technologies that help general prog rammers solve real problems.\n\nhttp://indico.cern.ch/conferenceDisplay.py ?confId=87133 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=87133 END:VEVENT BEGIN:VEVENT SUMMARY:Cells-as-a-Service: Enterprise-Grade Cloud Infrastructure Research at HP Laboratories DTSTART;VALUE=DATE-TIME:20100324T130000Z DTEND;VALUE=DATE-TIME:20100324T140000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-88316@cern.ch DESCRIPTION:\nThe presentation will describe research work in HP Laborator ies\non providing secure cloud infrastructure services. The talk will\ncov er some of the issues\, both legal and technical\, that enterprises\nand g overnment face when considering the use of cloud computing within\ntheir c ore IT systems. It will then describe a prototype system\,\nknown as Cells -as-a-Service\, that has been developed in HP Laboratories\nto provide som e of the guarantees that are required for widespread\nadoption by this cus tomer group.\n\nAbout the speaker\n\nPatrick Goldsack is a Distinguished T echnologist with Hewlett-Packard\nLaboratories in Bristol where he has bee n working since 1987. He has\nworked in a variety of areas ranging from fo rmal specification\,\ndistributed network monitoring and measurement techn ologies and most\nrecently distributed management systems. He is best know n for his\nwork on the open source distributed configuration and\nmanageme nt system "SmartFrog". Most recently he has been researching\nways of prov iding secure cloud infrastructure services for large\nenterprises and gove rnments. Patrick has an MA(oxon) in Mathematics from Oxford and an MSC in Electronic and Electrical Engineering.\n\n\nhttp://indico.cern.ch/conferen ceDisplay.py?confId=88316 LOCATION:CERN Kjell Johnsen Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=88316 END:VEVENT BEGIN:VEVENT SUMMARY:Concurrent Collections (CnC): A new approach to parallel programmi ng DTSTART;VALUE=DATE-TIME:20100507T073000Z DTEND;VALUE=DATE-TIME:20100507T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-91517@cern.ch DESCRIPTION:A common approach in designing parallel languages is to provid e some high level handles to manipulate the use of the parallel platform. This exposes some aspects of the target platform\, for example\, shared vs . distributed memory. It may expose some but not all types of parallelism\ , for example\, data parallelism but not task parallelism. This approach must find a balance between the desire to provide a simple view for the do main expert and provide sufficient power for tuning. This is hard for any given architecture and harder if the language is to apply to a range of a rchitectures. Either simplicity or power is lost.\n\n\nInstead of viewing the language design problem as one of providing the programmer with high l evel handles\, we view the problem as one of designing an interface. On on e side of this interface is the programmer (domain expert) who knows the a pplication but needs no knowledge of any aspects of the platform. On the o ther side of the interface is the performance expert (programmer or progra m) who demands maximal flexibility for optimizing the mapping to a wide ra nge of target platforms (parallel / serial\, shared / distributed\, homoge neous / heterogeneous\, etc.) but needs no knowledge of the domain. Concur rent Collections (CnC) is based on this separation of concerns. The talk w ill present CnC and its benefits.\n\nAbout the speaker\n\nKathleen Knobe h as focused throughout her career on parallelism especially compiler techno logy\, runtime system design and language design. She worked at Compass (a ka Massachusetts Computer Associates) from 1980 to 1991 designing compiler s for a wide range of parallel platforms for Thinking Machines\, MasPar\, Alliant\, Numerix\, and several government projects. In 1991 she decided t o finish her education. After graduating from MIT in 1997\, she joined Dig ital Equipment’s Cambridge Research Lab (CRL). She stayed through the DE C/Compaq/HP mergers and when CRL was acquired and absorbed by Intel. She c urrently works in the Software and Services Group / Technology Pathfinding and Innovation.\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=91 517 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=91517 END:VEVENT BEGIN:VEVENT SUMMARY:Software Aspects of IEEE Floating-Point Computations for Numerical Applications in High Energy Physics DTSTART;VALUE=DATE-TIME:20100511T140000Z DTEND;VALUE=DATE-TIME:20100511T150000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-91439@cern.ch DESCRIPTION:\nFloating-point computations are at the heart of much of the computing done in high energy physics. The correctness\, speed and accura cy of these computations are of paramount importance. The lack of any of these characteristics can mean the difference between new\, exciting physi cs and an embarrassing correction.\n\n\nThis talk will examine practical a spects of IEEE 754-2008 floating-point arithmetic as encountered in HEP ap plications. After describing the basic features of IEEE floating-point ar ithmetic\, the presentation will cover:\ncommon hardware implementations ( SSE\, x87)\ntechniques for improving the accuracy of summation\, multiplic ation and data interchange\ncompiler options for gcc and icc affecting flo ating-point operations\nhazards to be avoided\nAbout the speaker\n\nJeffre y M Arnold is a Senior Software Engineer in the Intel Compiler and Languag es group at Intel Corporation. He has been part of the Digital->Compaq->I ntel compiler organization for nearly 20 years\; part of that time\, he wo rked on both low- and high-level math libraries. Prior to that\, he was i n the VMS Engineering organization at Digital Equipment Corporation. In t he late 1980s\, Jeff spent 2½ years at CERN as part of the CERN/Digital J oint Project. In 2008\, he returned to CERN to spent 10 weeks working wit h CERN/openlab. Since that time\, he has returned to CERN multiple times to teach at openlab workshops and consult with various LHC experiments. J eff received his Ph.D. in physics from Case Western Reserve University.\n\ n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=91439 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=91439 END:VEVENT BEGIN:VEVENT SUMMARY: Building IaaS Clouds with OpenNebula and its Application to Grid Computing DTSTART;VALUE=DATE-TIME:20100602T073000Z DTEND;VALUE=DATE-TIME:20100602T083000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-96235@cern.ch DESCRIPTION:The aim of the seminar is to describe the innovations in cloud management brought by the OpenNebula Cloud Toolkit\, and its application to simplify and optimize the use and operation of cluster and grid computi ng infrastructures. OpenNebula is a widely used open-source Cloud manager that fits into existing data centers to build private\, public and hybrid Infrastructure-as-a-Service (IaaS) Clouds.Most of its innovative features have been developed to address requirements from business use cases in RES ERVOIR\, flagship of European research initiatives in virtualized infrastr uctures and cloud computing. The innovations comprise support for elastic multi-tier services\; flexible and scalable back-end for virtualization\, storage and networking management\; and support for Cloud federation and i nteroperability. OpenNebula brings to the grid and cluster ecosystem sever al benefits in terms of simplification\, flexibility\, maintainability\, o perational savings and resilience of the computing systems.The presentatio n ends with an introduction of the community and ecosystem that are evolvi ng around OpenNebula and the new European projects on cloud computing infr astructures that are using this innovative cloud technology. One of these projects is StratusLab\, which will incorporate cloud and virtualization i nnovation into existing Grid infrastructures by integrating cloud technolo gies and services within Grid sites\; and by enriching existing computing infrastructures with "Infrastructure as a Service" (IaaS) cloud-like deliv ery paradigms.About the speakerIgnacio M. Llorente\, Ph.D in Computer Scie nce (UCM) and Executive MBA (IE Business School)\, is a Full Professor in Computer Architecture and Technology\, and the Head of the Distributed Sys tems Architecture Research Group at Complutense University of Madrid. He h as 18 years of experience in research and development of advanced distribu ted computing and virtualization technologies\, architecture of large-scal e distributed infrastructures and resource provisioning platforms\, and ma nagement of international projects and initiatives on Grid and Cloud Compu ting.His current research interests are mainly in the area of Infrastructu re-as-a-Service (IaaS) Cloud Computing\, co-leading the research and devel opment of the OpenNebula Toolkit for Cloud Computing and coordinating the Activity on Management of Virtual Execution Environments in the RESERVOIR Project\, main EU-funded research initiative in virtualized infrastructure s and cloud computing. He founded and co-chaired the Open Grid Forum Worki ng Group on Open Cloud Computing Interface\; and participates in the Europ ean Cloud Computing Group of Experts.\n\nhttp://indico.cern.ch/conferenceD isplay.py?confId=96235 LOCATION:CERN Kjell Johnsen Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=96235 END:VEVENT BEGIN:VEVENT SUMMARY:Towards Energy-Centric Computing and Computer Architecture DTSTART;VALUE=DATE-TIME:20100603T090000Z DTEND;VALUE=DATE-TIME:20100603T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-95989@cern.ch DESCRIPTION:Technology forecasts indicate that device scaling will continu e well into the next decade.  Unfortunately\, it is becoming extremely di fficult to harness this increase in the number of transistors into perfor mance due to a number of technological\, circuit\, architectural\, methodo logical and  programming challenges.In this talk\, I will argue that the key emerging showstopper is power.  Voltage scaling as a means to maintai n a constant power envelope with an increase in transistor  numbers is hi tting diminishing returns. As such\, to continue riding the Moore's law we need to look  for drastic measures to cut power. This is definitely the case for server chips in future datacenters\, where abundant server paral lelism\, redundancy and 3D chip integration are likely to remove  program ming\, reliability and bandwidth hurdles\, leaving power as the only true limiter.I will present  results backing this argument based on validated models for future server chips and parameters  extracted from real commer cial workloads. Then I use these results to project future research  dire ctions for datacenter hardware and software.About the speakerBabak Falsafi is a Professor in the School of Computer and Communication Sciences at EP FL\, and an  Adjunct Professor of Electrical and Computer Engineering and Computer Science at Carnegie Mellon.  He is the founder and the directo r of  the Parallel Systems Architecture Laboratory (PARSA) at EPFL where he conducts research on architectural support for parallel programming\, resilient systems\, architectures to break the memory  wall\, and analyt ic and simulation tools for computer system performance evaluation.In 1999 \, in  collaboration with T. N. Vijaykumar he showed for the first time t hat\, contrary to conventional wisdom\, multiprocessors do not need rela xed memory consistency models (and the resulting convoluted programming in terfaces found and used in modern systems) to achieve high performance. He is a recipient of an NSF CAREER  award in 2000\, IBM Faculty Partnership Awards between 2001 and 2004\, and an Alfred P. Sloan  Research Fellowsh ip in 2004. He is a senior member of IEEE and ACM. \n\nhttp://indico.cern .ch/conferenceDisplay.py?confId=95989 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=95989 END:VEVENT BEGIN:VEVENT SUMMARY:Technical Computing - Improving Our Ability to Understand the Worl d DTSTART;VALUE=DATE-TIME:20100610T120000Z DTEND;VALUE=DATE-TIME:20100610T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-94727@cern.ch DESCRIPTION:Technical computing has become more and more important to scie nce\, engineering\, and business as digital technology has given us the ab ility to measure and monitor the world with ever-greater frequency and spe cificity. According to a recent report\, two years ago the amount of data that had been generated and stored was the equivalent of about 10 tons of reading material for every person on earth. Two years from now\, that's ex pected to have expanded more than six-fold.\n\nNew advances are providing the foundation for tools and applications that will make technical computi ng much more affordable and accessible. Soon\, a computer model that takes a team of advanced software programmers months to build and days to run w ill be within reach of any researcher\, engineer\, or financial analyst wo rking at the PC on their desk. And as technology continues to advance\, th ese models will become more and more complete and accurate in the way they represent the world. This will speed our ability to test new ideas\, impr ove processes\, and advance our understanding of systems.\n\nToday's trend s of large scale high performance computing\, the emergence of cloud compu ting\, and the parallel programming challenge and opportunity\, will trans form technical computing in a very significant way.  Harnessing these adv ances will fundamentally change our ability to process and analyze tremend ous volumes of data\, and our ability to build high fidelity models of the world around us.\n\nThis talk will discuss these significant trends in th e industry as well as the research and development Microsoft is pursuing i n this area with its newly announced \nTechnical Computing Initiative.\n\n About the speaker\nBill Hilf is the general manager of the Technical Compu ting group for Microsoft Corp.  In this role\, Bill oversees the world-wi de business for Technical Computing\, one of Microsoft's newest initiative s\, which focuses on solutions for a broad range of scientific and enginee ring challenges.  The Technical Computing division specializes in researc h and product development ranging from desktop applications to servers & c lusters\, to developer tools\, to cloud computing.Previously\, Bill was ge neral manager of Windows Server Marketing and Platform Strategy\, where he was responsible for global marketing of Microsoft's Windows Server produc ts and leading Microsoft's platform strategy and Open Source interoperabil ity efforts.Before joining Microsoft\, Hilf led IBM Corp.'s Linux/Open So urce Software technical strategy at a worldwide level. Prior to IBM\, Hilf held a variety of senior technical management and software architect posi tions for multiple Silicon Valley-based software companies. Hilf holds ove r ten patents in the distributed computing domain and is a graduate of Cha pman University Graduate School.\n\nhttp://indico.cern.ch/conferenceDispla y.py?confId=94727 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=94727 END:VEVENT BEGIN:VEVENT SUMMARY:Oracle JRockit: Advanced JVM Technologies DTSTART;VALUE=DATE-TIME:20100614T080000Z DTEND;VALUE=DATE-TIME:20100614T090000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-96236@cern.ch DESCRIPTION:In this seminar\, the Oracle Java product management team will introduce some of the advanced Java runtime technologies that are part of the\nOracle JRockit family. We will look at: the latest R28 release of t he Oracel JRockit\, including the new Flight Recorder (low overhead\, cont inuous recording) functionality that allows detailed profiling and analysi s of production Java applicationsJRockit Mission Control 4.0\, which provi des powerful visualization and analysis of JVM internals and application p erformance\, including detailed garbage collection statistics\, latency an alysis and memory leak detectionJRockit Real-Time\, which uses determinist ic garbage collection algorithms to optimize for extremely low-latency Jav a applications\, andJRockit Virtual Edition\, a special virtualization-awa re version of Oracle JRockit that allows the JVM to run directly on the Or acle VM (Xen) hypervisor without a guest OS.We will give hands-on demonstr ations of these technologies\, and there will be an opportunity for you to question Oracle's product specialists about our product plans and to dril l down into potential applications to the CERN Java community. The present ers will be: Tuva Palm from the Java Runtime Platform Group in Stockholm\, Erik Bergenholtz and Mark Prichard from the Java Platform Group in the US .\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=96236 LOCATION:CERN 513-1-027 URL:http://indico.cern.ch/conferenceDisplay.py?confId=96236 END:VEVENT BEGIN:VEVENT SUMMARY:A Distributed Cloud-Scale Execution Framework For Data DTSTART;VALUE=DATE-TIME:20100623T120000Z DTEND;VALUE=DATE-TIME:20100623T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-97661@cern.ch DESCRIPTION:The world of data that companies operate within has been drama tically transformed since the early days of data warehousing 20 to 30 year s ago. Data is the lifeblood of today's organizations\, and data growth is overwhelming traditional practices for managing and deriving value from i t. At the same time\, the cost of compute has continued to plummet\, follo wing Moore's Law\, and it is possible to build systems at 100 times better price/performance than just a few short years ago.The idea of the single EDW (Enterprise Data Warehouse) as the home for alldata simply hasn't come to fruition. The EDW plays an important role\, as a heavily controlled ce ntral repository that receives key distilled data from elsewhere in the co mpany. But increasingly it is locked down by compliance and process requir ements\, and it is simply not a place for data analysts and scientists to do innovative work on large\, new\, important datasets.Recognizing this im portant need\, Greenplum has developed\nGreenplum Chorus\,the world's firs t enterprise data cloud platform --a platform that leverages commodity com puting and private cloud computing ideas to break down the internalwalls b etween data silos and give analysts the freedom to derive insight from the m.This talk focuses on the opportunity for organizations to build an Enter prise Data Cloud as a net-new infrastructure to complement their EDW and r elated databases--one that is secure\, collaborative\, and data centric. A dditionally the technology behind Greenplum Chorus called Greenplum Data H ypervisor will be discussed. Greenplum Data Hypervisor is a distributed cl oud-scale execution framework that manages complex cross-database state\, data movement\, and orchestration in a fault-tolerant and self-healing man ner.About the speakerPrior to Greenplum\, Luke directed data center integr ation at High Performance Technologies Inc (HPTi)\, scaling the business a nd setting industry firsts in parallel computing subsequently adopted by I BM and Compaq. Luke's background includes 20+ years of innovations in comp uting technology ranging from supercomputing (#5 in top500 list) to medica l imaging systems. Previously he managed operations and R&D at groups with in Northrop Grumman Corporation. He holds an M.S. in Aeronautics and Astro nautics from Stanford University and a B.E. in Mathematics from Vanderbilt University.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=97661 LOCATION:CERN BE Auditorium Meyrin URL:http://indico.cern.ch/conferenceDisplay.py?confId=97661 END:VEVENT BEGIN:VEVENT SUMMARY:Building the Science SuperCloud: Applying the Lessons of Web 2.0 t o Science Data Processing DTSTART;VALUE=DATE-TIME:20100701T090000Z DTEND;VALUE=DATE-TIME:20100701T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-98857@cern.ch DESCRIPTION:The average size of a science dataset is increasing at an expo nential rate\, while advances in computing and storage capacity are increa sing at a linear rate. This raises the barrier-to-entry for many scientist s\, and frustrates efforts around open data and science collaboration. We will review some of the lessons learned while attempting to adapt Silicon Valley techniques to the challenges of NASA's science data. These lessons center on the use of commodity hardware\, open-source software\, and exper imental networking\, to achieve the goals of 'Cloud Computing' - elastic a nd self-provisioned use of shared resources.About the speakerJoshua McKent y has been pushing forward the causes of open science and open data at NAS A for the past two years. Prior to that he spent the last decade building teams and software\, in Silicon Valley\, Canada\, and the American Mid-Wes t.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=98857 LOCATION:CERN Kjell Johnsen Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=98857 END:VEVENT BEGIN:VEVENT SUMMARY:Fast parallel event reconstruction DTSTART;VALUE=DATE-TIME:20100706T120000Z DTEND;VALUE=DATE-TIME:20100706T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-99800@cern.ch DESCRIPTION:On-line processing of large data volumes produced in modern HE P experiments requires using maximum capabilities of modern and future man y-core CPU and GPU architectures.One of such powerful feature is a SIMD in struction set\, which allows packing several data items in one register an d to operate on all of them\, thus achievingmore operations per clock cycl e. Motivated by the idea of using the SIMD unit ofmodern processors\, the KF based track fit has been adapted for parallelism\, including memory opt imization\, numerical analysis\, vectorization with inline operator overlo ading\, and optimization using SDKs. The speed of the algorithm has been i ncreased in 120000 times with 0.1 ms/track\, running in parallel on 16  SPEs of a Cell Blade computer.  Running on a Nehalem CPU with 8 cores i t shows the processing speed of 52 ns/track using the Intel Threading Buil ding Blocks. The same KF algorithm running on an Nvidia GTX 280 in the C UDA frameworkprovides a plane throughput of 22 tracks/ms.In addition\, a many-core architecture code named Larrabee can be considered an interestin g platform to further scale the Kalman filter in the threading and vectori zation dimensions. Less architecture-dependent programming frameworks\, su ch as OpenCL and Intel Ct\,may also better support future changes in archi tecture. Thus\, for example\, the KF algorithm demonstrates a linear many- core scalability being implemented in the Intel Ct parallel language.The f ully SIMDized CA track finder of the future heavy-ion CBM experiment (FAIR /GSI) with the included SIMD KF track fit shows the full reconstruction ef ficiency of 92%. High energetic particles have the reconstruction efficien cy of 97%. The efficiency of low energetic tracks is 82% due to significan t multiple scattering in the detector material. The level of ghost tracks is only about 3%. The CA track finder demonstrates the maximum throughput of 150 centralor 1100 minimum bias events/s running on a Nehalem CPU wit h 8 cores. The strong many-core scalability of the CA track finder makes possible to keep the reconstruction at the event-level parallelism.More de tails on parallelism of the event reconstruction algorithms of the CBM\, a s well as ALICE and STAR experiments will be presented and discussed.A sho rt overview of the "Workshop for Future Challenges in Tracking and Trigger Concepts" (GSI\, Germany\, 07-11 June\, 2010) will be also given.\n\nhttp ://indico.cern.ch/conferenceDisplay.py?confId=99800 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=99800 END:VEVENT BEGIN:VEVENT SUMMARY:Moodle: How we built a community around open source software DTSTART;VALUE=DATE-TIME:20100809T120000Z DTEND;VALUE=DATE-TIME:20100809T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-102260@cern.ch DESCRIPTION:Martin will talk about the continuing growth of Moodle\n from an idea into a worldwide project helping millions of people. He'll talk ab out the factors that have contributed to Moodle's current success\, some t hat were integrated from the beginning of the project and some that were l earned along the way.About the speakerMartin Dougiamas is best known as th e founder of Moodle\, the popular free course management system used by mi llions of teachers around the world.As the executive director of Moodle Pt y Ltd in Perth\, Western Australia\, he leads the team of software develop ers at the heart of the Moodle project and the global network of 54 Moodle Partner service companies that help provide funding for this independent open source software project.Martin has a mixed academic background with p ost-graduate degrees in Computer Science and Education\, and continues to focus on researching how educators approach internet-based education. His major goal for the future is to improve the quality of education by encour aging social constructionist and collaborative practices within online lea rning communities.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=10 2260 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=102260 END:VEVENT BEGIN:VEVENT SUMMARY:Scalable Networking for Cloud Datacenters DTSTART;VALUE=DATE-TIME:20100921T090000Z DTEND;VALUE=DATE-TIME:20100921T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-103292@cern.ch DESCRIPTION:Andy will discuss the architectural evolution of Ethernet netw orks and switch architectures as they are being designed to address much l arger cloud networking applications that require predictable throughput an d latency.About the speakerAs Chief Development Officer\, Andy Bechtolshei m is responsible for the overall product development and technical directi on of Arista Networks.Previously Andy was a Founder and Chief System Archi tect at Sun Microsystems\, where most recently he was responsible for indu stry standard server architecture. Andy was also a Founder and President o f Granite Systems\, a Gigabit Ethernet startup acquired by Cisco Systems i n 1996. From 1996 until 2003 Andy served as VP/GM of the Gigabit Systems B usiness Unit at Cisco that developed the very successful Catalyst 4500 fam ily of switches. Andy was also a Founder and President of Kealia\, a next generation server company acquired by Sun in 2004.Andy received an M.S. in Computer Engineering from Carnegie Mellon University in 1976 and was a Ph .D. Student at Stanford University from 1977 until 1982..\n\nhttp://indico .cern.ch/conferenceDisplay.py?confId=103292 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=103292 END:VEVENT BEGIN:VEVENT SUMMARY:High quality software through collaborative development and testin g DTSTART;VALUE=DATE-TIME:20100923T080000Z DTEND;VALUE=DATE-TIME:20100923T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-99996@cern.ch DESCRIPTION:David Chadwick\, the test evangelist at IBM\, will share his v ision of software quality and collaborative development with pragmatic use cases and demonstrations of the Jazz platform.The event will start with a public presentation on IBM's perspective on today's software development quality through collaborative testing\, addressing points such as:Today's software quality\, why it matters?General overview of the IBM collaborativ e Jazz platform\, QA and test management-Demonstration of Rational Quality ManagerAutomatic functional testingCode quality assessmentCollaboration & management of the quality life-cycle with disparate teamsQuality metrics definitions and collection (including static code analysis)Requirement cov erageThe public part will be followed by an open dive-on session where fur ther details will be discussed.About the speakersDavid ChadwickRational Te sting Evangelist\, IBM Software Group\, USA\nDavid is currently a member o f the Jazz Jumpstart Team helping customers adapt the Jazz tools to their development and testing environments. For the previous six years\, he was a development and L3 support manager on the Rational Performance Tester te am.He has worked in the computer industry for 35 years including stints at Bell Labs\, Performance Awareness Corp.\, and Rational Software. For the last twenty years\, David has specialized in the development of performanc e testing technology. He has worked in product development and support\, t echnical services\, customer training\, and in pre-sales working with cust omers. He was also project leader and primary author for the IBM Redbook c alled "Using Rational Performance Tester version 7".His educational backgr ound includes two BS degrees (EE & CS) from NC State University\, an MSEE in Computer Engineering from Stanford University\, and completed PhD cours ework at Illinois Institute of Technology. \nRalph SchoonIBM Rational Jump start team Jazz-Collaboration\, Germany Ralph Schoon is currently a membe r of the Jazz Jumpstart Team helping customers adapt the Jazz tools to the ir development and testing environments.Prior to his current role Ralph wo rked as level 2 certified technical representative\, responsible for softw are development and application lifecycle management solutions in the indu strial sector. Since 2007 he specialized as subject matter expert and regi onal mentor on Jazz and Eclipse based solutions.Ralph has over 15 years of experience in software development and systems engineering and has worked as administrator\, developer and team lead. \nNicolas BerneyIBM Rational Account Manager\, Switzerland Nicolas Berney is currently responsible for IBM Rational in Suisse Romande and he is involved in helping customers to find the best solution for bettering their software development lifecycle .Prior to his current role Nicolas Berney was a senior Rational consultant for over 6 years in the area of change management and project management solutions to help customer to implement IBM products in Geneva and Zurich. Nicolas holds a master degree in Microelectronics from the engineering sch ool in Yverdon.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=99996 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=99996 END:VEVENT BEGIN:VEVENT SUMMARY:Design Patterns: establishing a discipline of parallel software en gineering DTSTART;VALUE=DATE-TIME:20100927T090000Z DTEND;VALUE=DATE-TIME:20100927T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-107382@cern.ch DESCRIPTION:Many core processors present us with a software challenge. We must turn our serial code into parallel code. To accomplish this wholesale transformation of our software ecosystem\, we must define established pra ctice is in parallel programming and then develop tools to support that pr actice. This leads to design patterns supported by frameworks optimized at runtime with advanced autotuning compilers. In this talk I provide an upd ate of my ongoing research with the ParLab at UC Berkeley to realize this vision. In particular\, I will describe our draft parallel pattern languag e\, our early experiments with software frameworks\, and the associated ru ntime optimization tools.About the speakerTim Mattson is a parallel progra mmer (Ph.D. Chemistry\, UCSC\, 1985). He does linear algebra\, finds oil\, shakes molecules\, solves differential equations\, and models electrons i n simple atomic systems. He has spent his career working with computer sci entists to make sure the needs of parallel applications programmers are me t.Tim has had the good fortune to work with brilliant people on truly grea t projects. Among these are:the first TFLOP computer (ASCI Red)\,the OpenM P API for shared memory programming\,the OpenCL programming language for h eterogeneous platforms\,programming Intel's first TFLOP chip (the 80 core research chip)\, andthe design and software architecture of Intel's 48 cor e research chipIn addition\, Tim has been working with a community of patt ern writers to create pattern languages for parallel programming. This wor k builds on his book "Patterns for Parallel Programming" (written with Pr ofessors Beverly Sanders and Berna Massingill and published by Addison Wes ley). \n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=107382 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=107382 END:VEVENT BEGIN:VEVENT SUMMARY:Simple\, intuitive and efficient parallel programming in Java DTSTART;VALUE=DATE-TIME:20101007T080000Z DTEND;VALUE=DATE-TIME:20101007T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-107951@cern.ch DESCRIPTION:\nI will present Ateji PX (http://www.ateji.com/px)\, a Java l anguage extension\nwith a set parallel programming primitives based on pi- calculus.\n\n\nUnlike API-based approaches\, we have retrofitted paralleli sm constructs at the\nlanguage level. This makes parallel programs simple and intuitive\, efficient\,\ncompatible with existing tools and practices\ , and provably correct.\n\nAteji PX has proven to be extremely easy to lea rn and use even by mainstream\nJava programmers without special training i n parallel programming. It also\nprovides a way to overcome many of the pr oblems inherent to the use of threads\nfor high level developers and HPC e xperts.\n\nThis seminar is divided into two parts. The first hour is a gen eral presentation of interest to all Java or HPC developers: I will presen t a survey of the new\nparallel programming constructs based on simple exa mples. You'll get a general\ngrasp of the language features and capabiliti es\, and a good understanding of howAteji PX compares to other technologie s such as Java threads\, ParallelArray\,\nOpenMP\, Cilk and MPI.\n\nI will cover the parallel programming styles\, including data-\, task-\, recursi ve-and speculative parallelism\, as well as data-flow\, stream programming \, the Actormodel and MapReduce.\n\nThe second hour of the seminar is targ eted to those who want to learn more aboutthis new parallel programming pa radigm. It will be based on hands-on demos and\ninteractive discussions\, answering all questions of the audience.\n\n\nparallelism\, as well\nas da ta-flow\, stream programming\, the Actor model and MapReduce.\n\nAteji PX is now available for multicore PCs and servers\, and soon for grid\,\nclou d and GPU accelerators. It is being adopted to boost computational intensi veJava applications such as risk calculations in finance\, protein folding in\nbioinformatics\, period analysis of satellite signals. Users report s ignificant\nimprovements in terms of development time\, training time\, co de quality\, ease of\nmaintenance\, and overall project development costs. \n\nAteji PX has been selected for presentation in the Disruptive Technolo gies\nexhibit at the SuperComputing 2010 conference in New Orleans.\n\nAbo ut the speaker\n\nAfter a PhD in computer science at INRIA\, Patrick Viry pursued an academic\ncarrier as a researcher at Kyoto University. He moved towards industry\, first bymanaging R&D projects for the japanese departm ent of technology (MITI)\, then as\na software architect at a major french software vendor.\n\nHe is an expert in both language engineering and comp utation models for parallelprogramming. When not coding\, he teaches japan ese language.\n\nHe founded Ateji in 2005 with the goal to make innovative language technologies\navailable to the community.\n\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=107951 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=107951 END:VEVENT BEGIN:VEVENT SUMMARY:EpiCollect: An open framework for mobile data collection DTSTART;VALUE=DATE-TIME:20101018T090000Z DTEND;VALUE=DATE-TIME:20101018T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-107398@cern.ch DESCRIPTION:\nEpiCollect is a generic\, free\, open data collection tool c onsisting of mobile phone software and a central web application. Users ca n collect and submit geotagged data forms (along with photos / videos ) to a central project website (hosted using Googles AppEngine) using suitable mobile phones (Android or iPhone). All data synchronised with the central website\, from multiple phones\, can then be viewed / charted / filtered using Google Maps / Earth or downloaded. Furthermore\, data can be reques ted and viewed/filtered from the project website directly on the phones us ing Google Maps.\n\n\nEpiCollect was initially developed for projects focu ssed on data gathering for Infectious Disease Epidemiology. However\, a fu lly generic version\, allowing anyone to setup their own projects online a nd focussed on data gathering in general has been released and this will b e the focus of the talk. Data can also be sent to any project web server\, not just those hosted using Google’s AppEngine. Further details about E piCollect can be found under "More Information" below.\n\nEpiCollect is cu rrently being used in a diverse set of projects\, for example describing A rchaeological dig sites\, monitoring invasive plant species and biodiversi ty\, disease monitoring across Africa\, and recording examples and locatio n of street art.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=1073 98 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=107398 END:VEVENT BEGIN:VEVENT SUMMARY:Next Generation Development Infrastructure: Maven\, M2Eclipse\, Ne xus & Hudson DTSTART;VALUE=DATE-TIME:20101108T100000Z DTEND;VALUE=DATE-TIME:20101108T111500Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-112200@cern.ch DESCRIPTION:\n All development organizations eventually converge on a set of tools to reduce costs\, lower onboarding time\, and leverage knowledge in strong communities to create standard processes. To this end we see in many organizations the emergence of a standard development stack consistin g of Maven\,M2Eclipse\, Nexus & Hudson. In this talk\, Jason van Zyl\, Fou nder of the Apache Maven project\, will discuss the future of Maven and sp ecifically Maven 3.x\, the rapidly approaching M2Eclipse 1.0 release\, the recent Nexus 1.8 release and roadmap\, and emerging tools such as Maven S hell and Polyglot Maven. Sonatype itself leverages this stack on a daily b asis and this discussion will focus not only on the tools individually\, b ut how they can work together to create a best practices approach to build ing and delivering your software in your organization.\n\n About the speak er\n\n Jason van Zyl is Chief Technical Officer and Founder of Sonatype\, and the founder of the Apache Maven project\, the Plexus IoC framework\, a nd the Apache Velocity project. Jason has over ten years of enterprise sof tware development experience. He founded Periapt\, a company that provided software infrastructure development services to Fortune 500 companies suc h as Toyota Corp.\, Bank of America\, and Coca-Cola Co. Previous to Periap t\, he worked as a Technology Architect at Compusense\, a world leader in sensory analysis and data research. Jason currently serves on the Apache M aven Project Management Committee. He has been involved with the Apache So ftware Foundation for seven years\, helped to found Codehaus\, a well resp ected incubation facility for open source community projects.\n\nhttp://in dico.cern.ch/conferenceDisplay.py?confId=112200 LOCATION:CERN BE Auditorium Meyrin URL:http://indico.cern.ch/conferenceDisplay.py?confId=112200 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Taipei CHEP 2010 Conference DTSTART;VALUE=DATE-TIME:20101119T083000Z DTEND;VALUE=DATE-TIME:20101119T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-105451@cern.ch DESCRIPTION:This year\, the 18th Computing in High Energy and Nuclear Phys ics (CHEP) conference has been held at Taipei\, Taiwan\, from 18-22 Octo ber 2010.CHEP conferences provide an international forum to exchange infor mation on computing experience and needs for the High Energy Physics and N uclear Physics communities\, and to review recent\, ongoing and future act ivities.The speakers at this Computing Seminar will present their views of the CHEP 2010 conference.\n\nhttp://indico.cern.ch/conferenceDisplay.py? confId=105451 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=105451 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Autumn 2010 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20101203T083000Z DTEND;VALUE=DATE-TIME:20101203T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-105449@cern.ch DESCRIPTION:\n The Autumn 2010 meeting of HEPiX was held at Cornell Univer sity\, New York USA on 1-5 November 2010. HEPiX is a global organizatio n of service managers and support staff providing computing facilities for the High Energy and Nuclear Physics community.\n\n The speakers will pres ent a summary of the subjects addressed during the meeting.\n\nhttp://indi co.cern.ch/conferenceDisplay.py?confId=105449 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=105449 END:VEVENT BEGIN:VEVENT SUMMARY:New trends in data analysis and visualization on the web DTSTART;VALUE=DATE-TIME:20110119T100000Z DTEND;VALUE=DATE-TIME:20110119T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-121591@cern.ch DESCRIPTION:\n Emerging technologies are democratizing  the power of serv er clusters to perform data analysis. At the same time new tools and new i deas are coming to data visualization on the web. Together this is creatin g a revolution on the Internet where data is combined with other data and visualized in amazing new ways. And this is now mainstream. The role this revolution has on science and how it is communicated is now starting to b ecome clearer with new trends like citizen science and crowd-sourcing. As a result\, researchers are exploiting a new kind of computation power\, h uman computation.\n\n In this seminar we will talk about the story behind some of these initiatives while presenting some new projects being develop ed by Vizzuality. From how to find new exoplanets\, to understand climate change or doing better nature conservation\, we will try to present you w ith as many ideas as possible so that you might find a way to use it on yo ur field or engage the general public in your work.\n\n About the speakers \n\n Javier de la Torre is a recognized expert on Biodiversity Informatics and geospatial technologies applied to Biodiversity. He is co-founder of Vizzuality\, a company working with top research institutes in the field o f biodiversity and conservation. He is professor of Data analysis and Vis ualization at the Master on Global Environmental Change at IE University.\ n\n Sergio Alvarez does research on data visualization\, information desig n and interface design with a special focus on geo-information\, map relat ed visualizations and geospatial data interaction. He is co-founder of Viz zuality\, and professor of Visualization at the Master on Global Environm ental Change at IE University.\n\nhttp://indico.cern.ch/conferenceDisplay. py?confId=121591 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=121591 END:VEVENT BEGIN:VEVENT SUMMARY:Advanced Features of Intel® C++ Composer XE for Linux DTSTART;VALUE=DATE-TIME:20110218T100000Z DTEND;VALUE=DATE-TIME:20110218T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-125167@cern.ch DESCRIPTION:\nPerformance-oriented features of the latest release of the I ntel®\nC/C++ compiler\, Intel® C++ Composer XE version 12.0\, will be pr esented. New\nand improved vectorization and auto-parallelization optimiza tion features\nfor Intel® Architecture-based processors\, including the I ntel® processor\ncode-named Sandy Bridge and the Intel® AVX architecture \, will be\nhighlighted. Intel® Cilk Plus and a new array notation\, tec hnologies for\nimplementing parallelism in applications\, will be presente d as well.\n\nAbout the speaker\n\nJeffrey M Arnold is a Senior Software E ngineer in the Intel\nCompiler and Languages group at Intel Corporation. H e has been part of the\nDigital->Compaq->Intel compiler organization for n early 20 years\; for part\nof that time\, he worked on both low- and high- level math libraries. Prior to\nthat\, he was in the VMS Engineering organ ization at Digital Equipment\nCorporation. In the late 1980s\, Jeff spent 2½ years at CERN as part of the\nCERN/Digital Joint Project. In 2008\, he returned to CERN to spend 10 weeks\nworking with CERN openlab. Since that time\, he has returned to CERN multiple\ntimes to teach at openlab worksh ops and consult with various LHC\nexperiments. Jeff received his Ph.D. in physics from Case Western Reserve\nUniversity (USA).\n\n\nhttp://indico.ce rn.ch/conferenceDisplay.py?confId=125167 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=125167 END:VEVENT BEGIN:VEVENT SUMMARY:The power of event-driven analytics in Large Scale Data Processing DTSTART;VALUE=DATE-TIME:20110224T090000Z DTEND;VALUE=DATE-TIME:20110224T113000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-123802@cern.ch DESCRIPTION:\n FeedZai is a software company specialized in creating high- ­‐throughput low-­‐latency data processing solutions. FeedZai develo ps a product called "FeedZai Pulse" for continuous event-­‐driven analy tics that makes application development easier for end users. It automatic ally calculates key performance indicators and baselines\, showing how cur rent performance differ from previous history\, creating timely business i ntelligence updated to the second. The tool does predictive analytics and trend analysis\, displaying data on real-­‐time web-­‐based graphics . In 2010 FeedZai won the European EBN Smart Entrepreneurship Competition\ , in the Digital Models category\, being considered one of the "top-­‐2 0 smart companies in Europe".\n\n The main objective of this seminar/works hop is to explore the topic for large-­‐scale data processing using Com plex Event Processing and\, in particular\, the possible uses of Pulse in the scope of the data processing needs of CERN. Pulse is available as open -­‐source and can be licensed both for non-­‐commercial and commerci al applications. FeedZai is interested in exploring possible synergies wit h CERN in high-­‐volume low-­‐latency data processing applications.\ n\n The seminar will be structured in two sessions\, the first one being aimed to expose the general scope of FeedZai's activities\, and the second focused on Pulse itself:\n\n 10:00-11:00 FeedZai and Large Scale Data Pr ocessing\n \n \n Introduction to FeedZai\n \n FeedZai Pulse and C omplex Event Processing\n \n Demonstration\n \n Use-Cases and Ap plications\n \n Conclusion and Q&A\n \n \n 11:00-11:15 Coffee break \n \n  \n \n 11:15-12:30 FeedZai Pulse Under the Hood\n \n \n A Fir st FeedZai Pulse Application\n \n PulseQL overview\n \n Defining KPIs and Baselines\n \n Conclusion and Q&A\n \n\n About the speaker s\n\n Nuno Sebastião is the CEO of FeedZai. Having worked for many years for the European Space Agency (ESA)\, he was responsible the overall desig n and development of Satellite Simulation Infrastructure of the agency. Ha ving left ESA to found FeedZai\, Nuno is currently responsible for the who le operations of the company. Nuno holds an M.Eng. in Informatics Engineer ing for the University of Coimbra\, and an MBA from the London Business Sc hool.\n\n Paulo Marques is the CTO of FeedZai\, being responsible for prod uct development. Paulo is an Assistant Professor at the University of Coim bra\, in the area of Distributed Data Processing\, and an Adjunct Associat ed Professor at Carnegie Mellon\, in the US. In the past Paulo lead a larg e number of projects for institutions like the ESA\, Microsoft Research\, SciSys\, Siemens\, among others\, being now fully dedicated to FeedZai. Pa ulo holds a Ph.D. in Distributed Systems from the University of Coimbra.\n \nhttp://indico.cern.ch/conferenceDisplay.py?confId=123802 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=123802 END:VEVENT BEGIN:VEVENT SUMMARY:Application Lifecycle Management - HP's approach DTSTART;VALUE=DATE-TIME:20110228T130000Z DTEND;VALUE=DATE-TIME:20110228T143000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-121443@cern.ch DESCRIPTION:\n Software applications are everywhere in today's organizatio ns. They now run the business\, and the business requires agility. But ev erything in this area is changing. Application teams that were once co-l ocated are now virtual\, global\, offshore and outsourced. Development pro cesses are evolving from sequential methods like waterfall\, to flexible  ones like Agile. And the technologies have changed from thin-client\, self -contained applications to complex composite applications that leverage cl oud services and RIA.\n\n As a result\, building and managing applications today is more complicated than ever. In order to manage this complexity\, IT organizations must optimize the entire application lifecycle\, from in ception to retirement. And managing it allows IT organizations to deliver applications quickly\, securely and reliably.\n\n With HP Application Lif ecycle Management (ALM) solutions\, you can improve the quality\, predicta bility and agility of your application lifecycle. The result is lower cost s\, higher-quality\, more agile business applications.\n\n The event will start with a presentation on HP's approach for Application Lifecycle Manag ement and continue with an End-to-End real life scenario demonstration usi ng the different HP solutions:\n\n Requirements Management\n \n Developm ent Management\n \n Quality Management\n \n Performance / Load and Funct ional (Gui & non Gui) tests automation\n \n Security Management\n\n To co nclude\, a business case based on a Swiss company facing problems with te st management will be presented describing:\n\n how an audit was conducte d for quality assurance\n \n test organization changes resulting from aud it\n \n implementation of Quality and Testing competency center\n \n sel ection of software for test management (Quality Center was selected)\n \n implementation and rollout of Quality Center\n \n performance testing ob jectives\n \n implementation of a Master Test Plan as standard for projec ts\n\n About the speakers\n\n Yuval Keren / HP Software\n Application Life cycle Management solution consultant (PreSales)\, Switzerland\n\n Currentl y responsible for HP Software in Suisse Romande in helping customers to fi nd the best solutions for improving their software quality.\n\n Prior to h is current role Yuval Keren was a Senior Consultant / Project Manager for over 10 years in the area of Software Quality Assurance\, helping customer s to define their testing strategy\, methodology\, tools implementation an d execution.\n\n Helen di Lallo / itecor\n Enterprise Testing Practice Lea der\n Responsible for business development and customer support for Testin g projects\n\n Helen works with a number of companies in Switzerland and F rance advising on Test Approach\, Methodology\, implementation of Testing Organizations to address the needs of manual\, automated and performance t esting. She has particular experience in the domain of ERP systems.\n\nhtt p://indico.cern.ch/conferenceDisplay.py?confId=121443 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=121443 END:VEVENT BEGIN:VEVENT SUMMARY:Attack Trends 2011 -or- Why Software Security DTSTART;VALUE=DATE-TIME:20110316T150000Z DTEND;VALUE=DATE-TIME:20110316T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-128587@cern.ch DESCRIPTION:\n In some sense\, software is the lifeblood of most modern co mplex systems. Software can fail\, but worse yet\, software can be intenti onally made to fail by attackers. Instead of defending our systems by isol ating them from the network (an impossible task)\, we must build security in from the beginning. Both social networking and mobile device security p rovide important security lessons that can inform a reasoned approach. Mod ern malicious code\, including the Zeus Trojan\, Stuxnet\, and other persi stent web threats\, is as sophisticated as it is insidious. And future tre nds in attacks are even more alarming\, leveraging rootkits\, multi-core a ttacks\, and hard-to-diagnose timing issues. Our sole recourse is software security. The good news is that we actually know what to do to build secu rity in. \n\n About the speaker\n\n Gary McGraw is the CTO of Cigital\, In c.\, a software security consulting firm with headquarters in the Washing ton\, D.C. area and offices throughout the world. He is a globally recogni zed authority on software security and the author of eight best selling bo oks on this topic. His titles include Software Security\, Exploiting Softw are\, Building Secure Software\, Java Security\, Exploiting Online Games\, and 6 other books\; and he is editor of the Addison-Wesley Software Secur ity series. Dr. McGraw has also written over 100 peer-reviewed scientific publications\, authors a monthly security column for informIT\, and is fre quently quoted in the press.\n\n Besides serving as a strategic counselor for top business and IT executives\, Gary is on the Advisory Boards of For tify Software (acquired by HP)\, Invincea\, Dasient\, and Raven White. His dual PhD is in Cognitive Science and Computer Science from Indiana Univer sity where he serves on the Dean¹s Advisory Council for the School of Inf ormatics. Gary served on the IEEE Computer Society Board of Governors and produces the monthly Silver Bullet Security Podcast for IEEE Security & Pr ivacy magazine (syndicated by informIT).\n\npodcast www.cigital.com/silver bullet\nblog www.cigital.com/justiceleague\nbook www.swsec.com\npersonal w ww.cigital.com/~gem\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId =128587 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=128587 END:VEVENT BEGIN:VEVENT SUMMARY:Social Computing DTSTART;VALUE=DATE-TIME:20110317T100000Z DTEND;VALUE=DATE-TIME:20110317T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-127484@cern.ch DESCRIPTION:\n The past decade has witnessed a momentous transformation in the way people interact with each other. Content is now co-produced\, sha red\, classified\, and rated by millions of people\, while attention has b ecome the ephemeral and valuable resource that everyone seeks to acquire. This talk will describe how social attention determines the production and consumption of content within both the scientific community and social me dia\, how its dynamics can be used to predict the future and the role that social media plays in setting the public agenda.\n\n About the speaker\n\ n Bernardo Huberman is a Senior HP Fellow and Director of the Social Compu ting Lab at Hewlett Packard Laboratories. He received his Ph.D. in Physics from the University of Pennsylvania\, and is currently a Consulting Profe ssor in the Department of Applied Physics at Stanford University. He origi nally worked in condensed matter physics\, ranging from superionic conduct ors to two-dimensional superfluids\, and made contributions to the theory of critical phenomena in low dimensional systems.\n\n In the field of info rmation sciences\, he predicted the existence of phase transitions in larg e scale distributed systems\, and developed an economics approach to the s olution of hard computational problems. Dr. Huberman is one of the creator s of the field of ecology of computation\, and editor of a book on the sub ject: “The Laws of the Web: Patterns in the Ecology of Information” (M IT Press\, 2001).\n\n You may read more about him here\n\nhttp://indico.ce rn.ch/conferenceDisplay.py?confId=127484 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=127484 END:VEVENT BEGIN:VEVENT SUMMARY:Electric Cloud's ElectricCommander and ElectricAccelerator DTSTART;VALUE=DATE-TIME:20110511T120000Z DTEND;VALUE=DATE-TIME:20110511T133000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-138841@cern.ch DESCRIPTION:\n Electric Cloud's solutions for software building and testin g have an impressive customer list in industry. Many rely on them for buil d and test automation\, tools that bring quality assurance and agile devel opment together. Andrew and Kev will present ElectricCommander\, a solutio n for building regular and continuous integration tests\, releases\, and t ry-out builds before patch submissions. It enables agile workflows with an easy yet powerful multi-platform configuration system. Electric Cloud als o offers ElectricAccelerator\, a tool to analyze and automatically optimiz e dependencies in builds. They will introduce it and explain how it can tr emendously reduce build times. They will show demonstrations of both produ cts.\n EVO access:\n Title: Electric Cloud's ElectricCommander and Electri cAccelerator\n Community: Universe\n Meeting Access Information:\n - Meeti ng URLhttp://evo.caltech.edu/evoNext/koala.jnlp?meeting=MIM9Ms2D2tDMDi9v9t Da9I\n - Phone Bridge\n ID: 337 3374\n Central European Summer Time (+0200 )\n Start 2011-05-11  13:30\n End   2011-05-11  16:30\n Japan Standard Time (+0900)\n Start 2011-05-11  20:30\n End   2011-05-11  23:30\n East ern Daylight Time (-0400)\n Start 2011-05-11  07:30\n End   2011-05-11  10:30\n Pacific Daylight Time (-0700)\n Start 2011-05-11  04:30\n End   2011-05-11  07:30\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId =138841 LOCATION:CERN 32-1-A24 URL:http://indico.cern.ch/conferenceDisplay.py?confId=138841 END:VEVENT BEGIN:VEVENT SUMMARY:The latest GCC release series and the special modes of its runtime C++ library DTSTART;VALUE=DATE-TIME:20110513T090000Z DTEND;VALUE=DATE-TIME:20110513T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-131493@cern.ch DESCRIPTION:\n Last March\, long awaited\, GCC 4.6.0 has been released by the GCC Project\, sporting a number of new features and non-trivial improv ements over the previous release series. In the first part of the talk\, I 'll try to provide an overview of the most interesting recent areas of act ivity in the project\, emphasizing what is being delivered in the 4.6.x se ries and what can be expected over the next months. I will especially focu s on the improved support for the upcoming C++0x ISO C++ standard\, both i n terms of core-language and library facilities (eg\, the first publicly a vailable implementation of the 'constexpr' features).\n\n In the second pa rt of the talk I will concentrate instead on the so-called special modes o f the library - debug-mode\, parallel-mode\, profile-mode - the last two p robably less known but also subject to rather intense efforts over the las t months. Open issues and difficulties\, in particular with respect to har monizing the work with the wider C++0x effort\, will not be concealed and the collaboration of all the interested parties solicited and strongly enc ouraged.\nAbout the speaker\n\nPaolo Carlini set up his first Linux system in 1995 and never looked\nback. Since 2002 co-maintains the GNU implement ation of the C++ runtime\nlibrary\, part of the GNU Compiler Collection\, and contributes also the\ndevelopment of the C++ front-end\, with a specia l focus on the new C++0x\nfeatures. In 2008 has joined the Oracle Linux te am\, and on behalf of it\nalso participates to the ISO C++ Standards activ ities and meetings\, as\nprincipal member of PL22.16.\n\nhttp://indico.cer n.ch/conferenceDisplay.py?confId=131493 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=131493 END:VEVENT BEGIN:VEVENT SUMMARY:Security Analysis of the Un-hackable Victorinox Secure Device DTSTART;VALUE=DATE-TIME:20110523T090000Z DTEND;VALUE=DATE-TIME:20110523T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-133976@cern.ch DESCRIPTION:\n In 2010\, Victorinox proposed a new USB storage device with data access secured with biometric fingerprints. According to the press r elease\, this device contains several uncommon protections such as custom fingerprint reader with blood oxygen and temperature sensor\, custom desig ned "MKI Schnuffi-X" chip and self-destruct mechanism if chip is physicall y attacked. In 2010 and 2011\, Victorinox organized several challenge cont ests (Las Vegas\, London\, Geneva) called "break the code". The winning pr ize was up to 250k$ to whom is able to hack the "un-hackable" Victorinox S ecure Device. Until today\, nobody has won the contest. In this talk\, we analyze the security of the "un-hackable" Victorinox Secure Device. We dem onstrate that this product does not contain all the security mechanism des cribed in the press release. We also show that the biometric authenticatio n can be easily bypassed with tools such as notepad.exe. Finally we explai n why it is very hard to win the contest even if the device is unsafe.\n\n About the speaker\n\n Martin Vuagnoux finished his PhD last year at Ecole Polytechnique Fédérale de Lausanne (EPFL) in the Security and Cryptogra phy Laboratory led by Prof. Serge Vaudenay (LASEC). Preivously\, he worked for 6 years as ethical hacker and founded two companies in Switzerland.\n \n His research domains are the cryptanalysis of symmetric cryptosystems a nd protocols (SSL/TLS\, WEP\, WPA\, RFID\, etc.)\, the automated discovery of vulnerabilities in software and operating systems\, the reverse engine ering of unknown protocols in embedded systems and the analysis and the ex ploitation of side-channel attacks such as compromising electromagnetic em anations of electronic devices.\n\nhttp://indico.cern.ch/conferenceDisplay .py?confId=133976 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=133976 END:VEVENT BEGIN:VEVENT SUMMARY:OpenCL and the quest for Performance Portability DTSTART;VALUE=DATE-TIME:20110617T090000Z DTEND;VALUE=DATE-TIME:20110617T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-140118@cern.ch DESCRIPTION:\n OpenCL is an important new standard for heterogeneous compu ting. With OpenCL\, a software developer can write a single program that r uns on everything from a cell phone to a node in a supercomputer. To reach its full potential\, however\, OpenCL needs to deliver more than portabil ity. It needs to deliver "performance portability". In this talk\, we will discuss the "performance portability" of OpenCL programs. Frankly\, achie ving good "performance portability" can be challenging with the current re lease of OpenCL\, so in addition to tips for writing "performance portable " code today\, we'll discuss possible future enhancements to OpenCL to imp rove "performance portability".\n\n About the speaker\n\n Tim Mattson is a parallel programmer (Ph.D. Chemistry\, UCSC\, 1985). The recurring theme of his career is to do whatever it takes to meet the needs of parallel app lication programmers.\n\n Tim has been with Intel since 1993 where he has worked with brilliant people on great projects including:\n\n the first T FLOP computer (ASCI Red)\,\n \n the OpenMP API for shared memory programm ing\,\n \n the OpenCL programming language for heterogeneous platforms\,\ n \n programming Intel's first TFLOP chip (the 80 core research chip)\, a nd\n \n Intel’s 48 core\, SCC research processor\n\n Tim has published extensively including the books Patterns for Parallel Programming (with Be verly Sanders and Berna Massingill\, Addison Wesley\, 2004) and An Introdu ction to Concurrency in Programming Languages (with Matthew J. Sottile and Craig E Rasmussen\, CRC Press\, 2009)\, and the OpenCL Programming Guide (with Aaftab Munshi\, Ben Gaster\, James Fung\, and Dan Ginsburg\, Addison Wesley\, summer 2011).\n\nEVO broadcast information\n\nCommunity: Univ erse\nPassword: OpenCL\n\nMeeting Access Information: here\n\nhttp://i ndico.cern.ch/conferenceDisplay.py?confId=140118 LOCATION:CERN Salle Bohr URL:http://indico.cern.ch/conferenceDisplay.py?confId=140118 END:VEVENT BEGIN:VEVENT SUMMARY:The Quest for green efficiency in Scientific Computing DTSTART;VALUE=DATE-TIME:20110623T090000Z DTEND;VALUE=DATE-TIME:20110623T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-139795@cern.ch DESCRIPTION:\nEnvironmental awareness and rising energy prices have made\n energy-efficiency an important issue for managing computing centers\nand o ther IT systems. So far\, research has mostly focused on hardware\nand inf rastructure aspects\, i.e. developing more efficient hardware or\noptimisi ng cooling but studying the system as a whole has received\nless interest. \n\n\nWe study the optimization of computing systems by viewing a computin g\nfacility as a production unit on which we apply well established\nopera tions management principles to run it more efficiently. Our\nresearch goal is to find ways to improve throughput and\nenergy-efficiency by schedulin g requests based on their estimated\nproperties and loads of all component s (CPU\, memory\, network) using\noperations management principles of opti mal loading.\n\n\nBecause of variations in jobs and the system behaviour\, it is not\npossible to find a fixed optimum utilisation level. Therefore\ , we\ndeveloped a fuzzy logic-based method that dynamically adapts\nparame ters. In this way\, it is possible to keep\, for example\, memory\nconsump tion stable with different workloads and this way achieving\nhigher throug hput and energy-efficiency than by using a traditional\nfixed number of jo bs or fixed memory threshold approaches. Our\npreliminary results in scien tific computing scenarios show that this\napproach can decrease energy con sumption by 50% while increasing\nthroughput up to 100% when compared to t he current standard practices\nin scientific computing.\n\nAbout the speak ers\n\nProf. Ari-Pekka Hameri is a professor of Operations Management at t he\nUniversity of Lausanne and the director of the Technology program at\n the Helsinki Institute of Physics. He received his PhD in Industrial\nMana gement from the Helsinki University of Technology in 1993. His\nresearch h as focused on efficient use of the full spectrum of\nproduction resources. He has published over 50 articles in\ninternational management and scienc e journals concerning industrial\nIT\, and management of production\, proj ects and supply chains. He has\nbeen involved with numerous international research and spin-off\nprojects related to industrial IT and production ma nagement.\n\n\nDr. Tapio Niemi is the leader of the GreenIT project at the Helsinki\nInstitute of Physics (HIP). He received his PhD in Computer Sci ence\nfrom the University of Tampere (2001). Prior joining HIP Technology\ nProgramme at CERN in 2004 he was a scientist and assistant professor\nat the University of Tampere. His research interests are green\ncomputing and data management. He has over 30 international\npublications in this field .\n\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=139795 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=139795 END:VEVENT BEGIN:VEVENT SUMMARY:Exascale Data Analysis DTSTART;VALUE=DATE-TIME:20110707T090000Z DTEND;VALUE=DATE-TIME:20110707T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-145169@cern.ch DESCRIPTION:\n Traditionaly\, the primary role of supercomputers was to cr eate data\, primarily for simulation applications. Due to usage and techno logy trends\, supercomputers are increasingly also used for data analysis. Some of this data is from simulations\, but there is also a rapidly incre asingly amount of real-world science and business data to be analyzed. We briefly overview Blue Gene and other current supercomputer architectures. We outline future architectures\, up to the Exascale supercomputers expect ed in the 2020 time frame. We focus on the data analysis challenges and op portunites\, especially those concerning Flash and other up-and-coming sto rage class memory.\n\n About the speakers\n\n Blake G. Fitch has been with IBM Research\, Yorktown Heights\, NY since 1987\, mainly pursuing interes ts in parallel systems. He joined the Scalable Parallel Systems Group in 1 990\, contributing to research and development that culminated in the IBM scalable parallel system (SP*) product. His research interests have focuse d on application frameworks and programming models suitable for production parallel computing environments. Practical application of this work inclu des contributions to the transputerbased control system for the 1994 IBM C MOS S/390* mainframes\, while at IBM Boeblingen\, Germany\, and to the arc hitecture of the 1996 IBM Automatic Fingerprint Identification System para llel application\, while at IBM Hursley\, UK. He joined the Blue Gene proj ect in 1999 as the application architect for Blue Matter\, a scalable mole cular dynamics package. From 2008 he is the technical lead for data storag e and analysis research in the Blue Gene program.\n\n Burkhard Steinmacher -Burow was a member of the ZEUS collaboration at DESY from 1988 to 1996. H is PhD thesis measured the total photon-proton cross section. In the early 1990's he led the development of Funnel\, the ZEUS production facilty for simulated events\, using hundreds of workstations around the world. He ha s been with the Blue Gene project since 2001\, primarily contributing to t he network across nodes and to the memory system within a node. Initially based at IBM Research\, Yorktown Heights\, NY\, in 2004 he moved to IBM Re search and Development Germany in Boeblingen\, where he also is the techni cal lead of the local Blue Gene team.\n\nhttp://indico.cern.ch/conferenceD isplay.py?confId=145169 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=145169 END:VEVENT BEGIN:VEVENT SUMMARY:Designing for Communities DTSTART;VALUE=DATE-TIME:20110901T090000Z DTEND;VALUE=DATE-TIME:20110901T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-150063@cern.ch DESCRIPTION:\n In 2008\, Mark Boulton's small design firm won the project to redesign drupal.org home to the Drupal open source content management project and its community of over 400\,000 users and developers. Designing for open source communities has its challenges: it is a developer-centric environment where discussion\, debate and consensus rules and where desig n processes are often viewed sceptically. To help dispel some misconceptio ns behind the design process\, Mark's team worked completely openly with t he Drupal community - and broader the web design community - throughout t he project.\n\n This talk will walk through some of the challenges and sol utions for designing for large communities. Mark will share the successes - and the horror stories - of how a traditional design process was challe nged every step of the way and how the Drupal community became an integral part of that new process.\n\n Mark and his team will be working with the Communication Group on an open project to redesign a new core website for the CERN community.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=1 50063 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=150063 END:VEVENT BEGIN:VEVENT SUMMARY:The Power of Hybridization DTSTART;VALUE=DATE-TIME:20111014T090000Z DTEND;VALUE=DATE-TIME:20111014T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-157945@cern.ch DESCRIPTION:\nProgramming languages always seem to do some things well but not others: Python punts when it comes to user interfaces\, Java’s arti ficial complexity prevents rapid development and produces tangles\, and it will be awhile before we see benefits from C++ concurrency work. The cogn itive load of languages and their blind spots increases the cost of experi mentation\, impeding your ability to fail fast and iterate. If you use a s ingle language to solve your problem\, you are binding yourself to the wor ldview limitations and the mistakes made by the creator of that language.\ n\nConsider increasing your wiggle room by crossing language boundaries\, complementing a language that is powerful in one area with a different lan guage powerful in another. Language hybridization can speed development to quickly discover your real problems\, giving you more time to fix them.\n \nAfter making a case for hybridizing your thinking in general\, I will pr esent a number of simple examples\; first showing the benefits of using ot her languages with multiprocessing in Python and Actors in Scala\, then hy bridization creating a Go language JSON-RPC server and a Python client\, a nd finally a Python web server with a web client using CoffeeScript\, jQue ry and Ajax. All examples are kept small so that the syntax of each new la nguage can be explained.\n\n\n About the speaker\n\n Bruce Eckel is the au thor of Thinking in Java (Prentice-Hall\, 1998\, 2nd Edition\, 2000\, 3rd Edition\, 2003\, 4th Edition\, 2006)\, the Hands-On Java Seminar CD ROM (a vailable on the Web site)\, Thinking in C++ (PH 1995\; 2nd edition 2000\, Volume 2 with Chuck Allison\, 2003)\, C++ Inside & Out (Osborne/McGraw-Hil l 1993)\, and First Steps in Flex (with James Ward\, 2008) among others. H e's given hundreds of presentations throughout the world\, published over 150 articles in numerous magazines\, was a founding member of the ANSI/ISO C++ committee and speaks regularly at conferences. He provides public and private training and consulting in programming languages and software sys tem design.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=157945 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=157945 END:VEVENT BEGIN:VEVENT SUMMARY:The rise of HPC accelerators: towards a common vision for a petasc ale future DTSTART;VALUE=DATE-TIME:20111104T100000Z DTEND;VALUE=DATE-TIME:20111104T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-155860@cern.ch DESCRIPTION:\n Nowadays new exciting scientific discoveries are mainly dri ven by large challenging simulations. An analysis of the trends in High Pe rformance Computing clearly show that we hit several barriers (CPU frequen cy\, power consumption\, technological limits\, limitations of the present paradigms) that we cannot easily overcome. In this context\, accelerators became the concrete alternative to increase the compute capabilities of t he deployed HPC clusters inside Universities and research centers across E urope. Within the EC funded "Partnership of Advanced Computing in Europe" (PRACE) project\, several actions has been taken and will be taken to enab le community codes to exploit accelerators in modern HPC architectures. In this talk\, the vision and the strategy adopted by the PRACE project will be presented\, focusing on new HPC programming model and paradigm. Accele rators are a fundamental piece to innovate in this direction\, from both t he hardware and the software point of view. This work started during the P RACE Preparatory Phase (2008-2010) and has been carried on during the PRAC E First Implementation Phase (1IP\, 2010-2012).\n\n About ICHEC\n\n The Ir ish Centre for High-end Computing (ICHEC) leads the Sub-Task "Accelerator" within the Task "Programming Techniques for High Performance Applications " inside the Work Package "Enabling Petascale Applications: Efficient use of Tier-0 systems". The main porting activities performed until now concer n the adoption of GP-GPU technologies (mainly NVIDIA® CUDA) in real appli cations. ICHEC is directly involved in the porting of Quantum ESPRESSO and EC-EARTH/IFS\, a material science and a weather forecast code. We expect in the Second Implementation Phase (2IP\, 2011-2013) to explore and assess the adoption of OpenCL to target a wide portfolio of accelerators like AT I® GPU\, AMD® APU or Intel® Many Integrated Core Architecture. Portabil ity\, reliability and maintenance of future codes will be fundamental aspe cts to evaluate and adopt new hardware.\n\n Notice: This presentation cont ains personal opinions\, forecasts and forward-looking statements of the a uthors and not necessarily the opinions of any other person or organizatio n mentioned. For official communications and disclosures please refer to P RACE or ICHEC websites.\n\nhttp://indico.cern.ch/conferenceDisplay.py?conf Id=155860 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=155860 END:VEVENT BEGIN:VEVENT SUMMARY:The Oracle NoSQL Database DTSTART;VALUE=DATE-TIME:20111108T100000Z DTEND;VALUE=DATE-TIME:20111108T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-159716@cern.ch DESCRIPTION:\nThe Oracle NoSQL Database is a distributed key-value databas e. It is\ndesigned to provide highly reliable\, scalable and available dat a storage\nacross a configurable set of systems that function as storage n odes.\n\nData is stored as key-value pairs\, which are written to particul ar storage\nnode(s)\, based on the hashed value of the primary key. Storag e nodes are\nreplicated to ensure high availability\, rapid failover in th e event of a\nnode failure and optimal load balancing of queries. Custome r applications\nare written using an easy-to-use Java API to read and writ e data. The NoSQL\nDatabase links with the customer application\, providin g access to the data\nvia the appropriate storage node for the requested k ey-value.\n\nThis talk will discuss the architecture and technical details of Oracle\nNoSQL Database.\n\nAbout the speaker\n\nCharles Lamb is a deve loper on Oracle NoSQL Database as well as Berkeley DB\nJava Edition.\n\nht tp://indico.cern.ch/conferenceDisplay.py?confId=159716 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=159716 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Autumn 2011 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20111111T090000Z DTEND;VALUE=DATE-TIME:20111111T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-162378@cern.ch DESCRIPTION:\n The Autumn 2011 meeting of HEPiX was held at TRIUMF\, Vanc ouver\, Canada on 24-28 October 2011. HEPiX is a global organization of service managers and support staff providing computing facilities for the High Energy and Nuclear Physics community.\n\n The speakers will present a summary of the subjects addressed during the meeting.\n\nhttp://indico.c ern.ch/conferenceDisplay.py?confId=162378 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=162378 END:VEVENT BEGIN:VEVENT SUMMARY:Software Defects\, Scientific Computation and the Scientific Metho d DTSTART;VALUE=DATE-TIME:20111214T100000Z DTEND;VALUE=DATE-TIME:20111214T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-161508@cern.ch DESCRIPTION:\n Computation has rapidly grown in the last 50 years so that in many scientific areas it is the dominant partner in the practice of sci ence. Unfortunately\, unlike the experimental sciences\, it does not adher e well to the principles of the scientific method as espoused by\, for exa mple\, the philosopher Karl Popper. Such principles are built around the n otions of deniability and reproducibility.\n\n Although much research effo rt has been spent on measuring the density of software defects\, much less has been spent on the more difficult problem of measuring their effect on the output of a program. This talk explores these issues with numerous e xamples suggesting how this situation might be improved to match the deman ds of modern science.\n\n Finally it develops a theoretical model based on an amalgam of statistical mechanics and Hartley/Shannon information theor y which suggests that software systems have strong implementation independ ent behaviour and supports the widely observed phenomenon that defects clu ster.\n\n About the speaker\n\n Les Hatton is part-time Professor of Foren sic Software Engineering at Kingston University\, London. He graduated in mathematics from King's College Cambridge and Manchester Universities shor tly after the last Ice Age. After being awarded the Conrad Schlumberger Pr ize for computational geophysics in 1987\, he was lured into the dark worl d of software failure and never came back.\n\n He has wide interests and h as just published his latest book "Email Forensics: Eliminating Spam\, Sca ms and Phishing."\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=161 508 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=161508 END:VEVENT BEGIN:VEVENT SUMMARY:SlipStream: automated provisioning and continuous deployment in th e cloud DTSTART;VALUE=DATE-TIME:20120201T100000Z DTEND;VALUE=DATE-TIME:20120201T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-175046@cern.ch DESCRIPTION:\n Cloud technology is now everywhere. Beyond the hype\, it pr ovides a real opportunity to improve the engineering of software systems. Lately the DevOps movement has also gain momentum\, which take an agile ap proach at bringing developers and system administrators closer together to better engineer software systems.\n\n In this context\, this presentation focuses on new tools for exploiting cloud services (private and public) i n order to create a continuous flow between software commits and fully dep loyed and configured software systems\, automatically and on-demand. To il lustrate this\, we present SlipStream and StratusLab.\n\n SlipStream is a new product developed by SixSq\, able to create virtual machines and orche strate multi-machine deployments.  SlipStream started from an idea develo ped in the context of the ETICS project\, led by CERN. StratusLab is an op en-source IaaS distribution\, able to create public and private clouds.\n\ n This presentation will also describe a case study where SlipStream depl oys an entire Mission Control System for the Operations Center of the Euro pean Space Agency.  We also describe how SlipStream can be integrated wit h management tools\, such as Puppet and Chef.\n\n About the speaker\n\n Ma rc-Elian Bégin is co-founder of SixSq. At SixSq he is the lead developer of the SlipStream product\, leads the Integration and Test Work Package of the StratusLab FP7 project\, delivers regular agile training and is coach for different customers\, including the Operation Centre of ESA (ESOC). H e worked at CERN from 2004 to 2008 in the IT Department\, as developer and technical lead on the EGEE and ETICS projects. Before joining CERN\, he w orked for a decade in the space industry\, contributing to the delivery of several software systems\, both for the space agencies and industry in Ca nada and Europe. Marc-Elian is a regular speaker at international events\, such as Agile2011\, XP2011 and DevOpsDays. In Helix Nebula\, he leads the Architecture and Technical Group\, responsible for short and longer term technical solutions for the cloud collaboration.\n\nhttp://indico.cern.ch/ conferenceDisplay.py?confId=175046 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=175046 END:VEVENT BEGIN:VEVENT SUMMARY:Many-core technologies: The move to energy-efficient\, high-throug hput x86 computing (TFLOPS on a chip) DTSTART;VALUE=DATE-TIME:20120206T150000Z DTEND;VALUE=DATE-TIME:20120206T160000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-171978@cern.ch DESCRIPTION:\n With Moore's Law alive and well\, more and more parallelism is introduced into all computing platforms at all levels of integration a nd programming to achieve higher performance and energy efficiency. Especi ally in the area of High-Performance Computing (HPC) users can entertain a combination of different hardware and software parallel architectures and programming environments. Those technologies range from vectorization and SIMD computation over shared memory multi-threading (e.g. OpenMP) to dist ributed memory message passing (e.g. MPI) on cluster systems. We will disc uss HPC industry trends and Intel's approach to it from processor/system a rchitectures and research activities to hardware and software tools techno logies. This includes the recently announced new Intel(r) Many Integrated Core (MIC) architecture for highly-parallel workloads and general purpose\ , energy efficient TFLOPS performance\, some of its architectural features and its programming environment. At the end we will have a brief look at Exa-Scale computing\, its challenges and opportunities.\n\n About the spea ker\n\n Dr. Herbert Cornelius is WW HPC Solution Architect at Intel with f ocus on technical\, high-performance computing (HPC) and many-core computi ng. Before he was Engineering Manager in Intel's Cluster Software & Techno logies group in EMEA\, working on scalable parallel computing hardware & s oftware solutions based on vectorization\, multi-threading and message-pas sing utilizing multi-core/multi-processor cluster platforms. Before joinin g Intel\, he served as Manager High-End Computing Europe at Fujitsu and wo rked at Cray Research from 1983 to 1990. He received a Ph.D. degree in Mat hematics and Diploma degree in Mathematics and Informatics from Technical University of Berlin\, Germany.\n\n  \n\nhttp://indico.cern.ch/conference Display.py?confId=171978 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=171978 END:VEVENT BEGIN:VEVENT SUMMARY:Dataflow Computing for Data-Intensive Applications DTSTART;VALUE=DATE-TIME:20120329T090000Z DTEND;VALUE=DATE-TIME:20120329T103000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-181589@cern.ch DESCRIPTION:\n Maxeler Technologies provides Maximum Performance Computing (MPC) based on a dataflow model of computation. Combined with a multi-di sciplinary approach\, these dataflow solutions are in production\, reduci ng power consumption and data center space for a wide range of applicati ons. Typically\, our dataflow solutions utilize thousands of arithmetic units on a chip to outperform top end microprocessors by 20x-40x in compu tations per cubic foot and computations per Watt\, and our analysis sugg ests that there is strong potential to do a lot better in future. Dataflo w is particularly suitable for highly demanding applications such as: fin ite element and finite difference PDE solvers on structured and unstructu red grids\, Monte Carlo methods\, multi-dimensional optimization problems \, real-time data processing\, and even sparse matrix solvers. At the co re of the approach\, we are building computers to match the problem\, ra ther than optimizing algorithms to standard microprocessors.\n\n Given the success in Earth Sciences\, Quantitative Finance and Electronic Trading  and the availability of a user-friendly dataflow programming environment \, we believe that this novel approach to computing could potentially lea d to scientific breakthroughs in data-intensive branches of science.\n I n spite of their spectacular speed\, FPGAs\, GPGPUs and also multicore arc hitectures obey to a restrictive computation paradigm that makes less attr active the numerical solution of many algebraic problems. Hence the real c hallenge for the developer is the reduction of a mathematical model to a s equence of computational tasks that perfectly fit the paradigm supported b y these extreme architectures. In a second part\, we illustrate this con cept for time imaging algorithms of general use in oil industry.\n  \n\n About the speakers\n\n Dr. Oskar Mencer\, Maxeler Technologies\n\n Prior t o founding Maxeler\, Oskar was Member of Technical Staff at the Computing Sciences Center at Bell Labs in Murray Hill\, leading the effort in "Str eam Computing". He joined Bell Labs after receiving a PhD from Stanford U niversity. Besides driving Maximum Performance Computing (MPC) at Maxeler \, Oskar is Consulting Professor in Geophysics at Stanford University and he is also affiliated with the Computing Department at Imperial College London\, having received two Best Paper Awards\, an Imperial College Rese arch Excellence Award in 2007 and a Special Award from Com.sult in 2012 f or "revolutionising the world of computers".\n\n Dr. Ernesto Bonomi\, CRS 4\n\n Director at CRS4 of three highly-motivated research groups (Imaging and Numerical Geophysics\, Environmental Sciences\, and Process Engineerin g and Combustion) which include about 20 research staff\, developing and a pplying numerical simulation models and data analysis tools for earth expl oration and seismic imaging\, hydrology and territorial planning\, meteoro logy\, and clean combustion processes.\n\n Head of the developers? team pr oviding innovative industrial acoustic imaging tools based on original mat hematical developments and on efficient data-driven parallel algorithms gi ving rise to the implementation of very-large-scale 3D numerical applicati ons used routinely in production on HPC platforms and\, more recently\, on acceleration hardware (FPGA and GPU).\n\nhttp://indico.cern.ch/conference Display.py?confId=181589 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=181589 END:VEVENT BEGIN:VEVENT SUMMARY:Evolution of the CERN Computing Infrastructure DTSTART;VALUE=DATE-TIME:20120401T120000Z DTEND;VALUE=DATE-TIME:20120401T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-185636@cern.ch DESCRIPTION:\n Large scale computing has evolved since the present Compute r Centre infrastructure and tools were designed to manage the first years of LHC data taking. Today we can see the scaling limitations of some of th em and questions are raised over the role of traditional batch systems in the future\; many tools and technologies exist in the public domain to man age facilities at large scale. This\, together with changes to the way we will provide computing in the future\, lead us to reassess the tools used to manage the infrastructure and the way in which services are managed.\n\ n The CERN Computer Centre will soon be extended with a remote facility\, and virtualization and other technologies allow us to provide services in new ways. This talk will describe the work being done in IT to modernize t he tools and infrastructure to allow us to manage machines and services at significantly larger scales than today\,  how we will make use of virtua lization to more effectively use the computing resources\, and the implica tions of the addition of a remote extension to the Computer Centre. It wil l also cover how we see cloud services playing a role in the future at CER N and how these can be seen as complementary to the grid and traditional s ervices.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=185636 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=185636 END:VEVENT BEGIN:VEVENT SUMMARY:Models\, DSLs and Code Generation DTSTART;VALUE=DATE-TIME:20120424T090000Z DTEND;VALUE=DATE-TIME:20120424T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-171981@cern.ch DESCRIPTION:\n Models are abstractions of the real world. They express onl y those aspects of the target system important for the model's purpose\, s uch as analysis or synthesis. Domain Specific Languages (DSLs) are languag es tailored to expressing models in a specific problem domain\; they expre ss those models concisely and in a way that makes the implied semantics "o bvious". Finally\, code generation is the way to synthesize executable cod e from domain-specific models. In this seminar I will introduce the basic concepts and illustrate some of the benefits of using models\, DSLs and co de generation in software development\, show a couple of example languages and introduce two of the most powerful tools in this area: Eclipse Xtext and JetBrains MPS.\n\n About the speaker\n\n Markus Völter works as an in dependent researcher\, consultant and coach for itemis AG in Stuttgart\, G ermany. His focus is on software architecture\, model-driven software deve lopment and domain specific languages as well as on product line engineeri ng. Markus also regularly writes (articles\, patterns\, books) and speaks (trainings\, conferences) on those subjects. Contact him via voelter at ac m dot org or www.voelter.de\n\nhttp://indico.cern.ch/conferenceDisplay.py ?confId=171981 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=171981 END:VEVENT BEGIN:VEVENT SUMMARY:Report on the Spring 2012 HEPiX Meeting DTSTART;VALUE=DATE-TIME:20120511T080000Z DTEND;VALUE=DATE-TIME:20120511T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-165625@cern.ch DESCRIPTION:\n HEPiX is a global organization of service managers and supp ort staff providing computing facilities for the High Energy and Nuclear P hysics community.\n\n The speakers will present a summary of the subjects addressed during the meeting.\n\nhttp://indico.cern.ch/conferenceDisplay.p y?confId=165625 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=165625 END:VEVENT BEGIN:VEVENT SUMMARY:Writing parallel programs that work DTSTART;VALUE=DATE-TIME:20120608T120000Z DTEND;VALUE=DATE-TIME:20120608T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-191117@cern.ch DESCRIPTION:\n Serial algorithms typically run inefficiently on parallel m achines. This may sound like an obvious statement\, but it is the root cau se of why parallel programming is considered to be difficult. The current state of the computer industry is still that almost all programs in existe nce are serial. This talk will describe the techniques used in the Intel P arallel Studio to provide a developer with the tools necessary to understa nd the behaviors and limitations of the existing serial programs. Once the limitations are known the developer can refactor the algorithms and reana lyze the resulting programs with the tools in the Intel Parallel Studio to create parallel programs that work.\n\n About the speaker\n\n Paul Peters en is a Sr. Principal Engineer in the Software and Solutions Group (SSG) a t Intel. He received a Ph.D. degree in Computer Science from the Universit y of Illinois in 1993. After UIUC\, he was employed at Kuck and Associates \, Inc. (KAI) working on auto-parallelizing compiler (KAP)\, and was invol ved in the early definition and implementations of OpenMP. While at KAI\, he developed the Assure line of parallelization/correctness products\, for Fortran\, C++ and Java. In 2000\, Intel Corporation acquired KAI\, and he joined the software tools group. At Intel\, he worked with the tools grou p to create the Thread Checker products\, which evolved into the Inspector and Advisor components of the Intel® Parallel Studio.  Inspector uses d ynamic binary instrumentation to detect memory and concurrency bugs\, and Advisor uses similar techniques along with performance measurement and mod eling to assist developers in transforming existing serial applications to be ready for parallel execution.\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=191117 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=191117 END:VEVENT BEGIN:VEVENT SUMMARY:Building\, Testing\, and Deploying Software in a Cross-Platform De velopment Environment DTSTART;VALUE=DATE-TIME:20120619T090000Z DTEND;VALUE=DATE-TIME:20120619T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-194858@cern.ch DESCRIPTION:\n CMake has been in development since 1999\, and has been use d on several large open source projects such as ITK\, VTK\, ParaView\, VXL \, and CMake itself. Further\, KDE\, one of the largest OSS projects has a dopted it\, demonstrating that CMake is capable of successfully supporting complex and large software systems. Hence its usage is growing rapidly w ith thousands of daily downloads and inclusion in several Linux distributi ons.\n\n In addition to building software\, CMake provides a testing clien t (CTest) that integrates with the web-based CDash testing server. This se rver creates dashboards that build a snapshot of the software at a given t ime. This is critical to cross-platform development since often a change o n one platform fails to compile on another one. The testing system provide s for nightly builds which use a copy of the software at a specific time e ach night\, experimental tests that can be used to share build results wit h other developers before committing source code\, and continuous build re sults that test the build each time files are committed to the source cont rol system.\n\n Once the software is built and tested\, the CPack tool can be used to package the software. CPack works similar to CMake in that it generates package information for native packaging tools. NSIS\, RPM\, OSX packages self extracting tar.gz\, tar.gz\, tar.zip can all be created. CP ack information is included as a simple extension to the CMake build files .\n\n This seminar will cover the history\, basic usage and the future of CMake\, CTest\, CDash and CPack in the context of an integrated developmen t environment.\n\n About the speaker\n\n Julien Jomier is directing Kitwar e's European subsidiary in Lyon\, France (Kitware SAS). Prior to joining K itware\, Mr. Jomier was a Faculty Research Lecturer of Radiology at the Un iversity of North Carolina.\n\n Julien has been leading the MIDAS project\ , a system for collecting\, processing\, and distributing massive collecti ons of data. He is also the main architect of CDash\, an open-source\, dis tributed\, software quality system companion of CMake and CTest.\n\n Julie n received both his B.S. and M.S in Electrical Engineering and Information Processing in 2002 from the ESCPE-Lyon (France) and an M.S. in Computer S cience from The University of North Carolina at Chapel Hill (UNC) in 2003. He worked on a variety of projects in the areas of parallel and distribut ed computing\, mobile computing\, image processing and visualization.\n\nh ttp://indico.cern.ch/conferenceDisplay.py?confId=194858 LOCATION:CERN 32-1-A24 URL:http://indico.cern.ch/conferenceDisplay.py?confId=194858 END:VEVENT BEGIN:VEVENT SUMMARY:Trends in Database Research DTSTART;VALUE=DATE-TIME:20120620T120000Z DTEND;VALUE=DATE-TIME:20120620T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-192081@cern.ch DESCRIPTION:\n In recent years we have witnessed a Cambrian explosion in d atabase research. Column-stores\, NoSQL\, MapReduce have broken many of th e assumptions present in traditional (relational) database systems. In thi s talk\, we start by looking briefly at the early days of database researc h\, making parallels with the current state of affairs and drawing importa nt lessons. We will then discuss a few selected database research topics\, including some lesser-known techniques\, discussing their impact given co mputer architecture trends as well as new application requirements. In par ticular\, we will show how a few important facts about modern hardware des ign are conditioning the architecture of database systems\, speculating ho w future database systems are expected to come about.\n\n About the speake r\n\n Miguel Branco is a postdoctoral researcher at the Data-Intensive App lications and Systems Lab at EPFL. Between 2003 and 2009\, he worked at CE RN developing the distributed data management system of the ATLAS Experime nt. At EPFL\, his work focuses on adapting data management technology to c omputer architecture trends\, where he has co-developed techniques to effi ciently run transactional database systems in heterogeneous multisocket/mu lticore machines\, as well as novel database architectures for "in situ" q uery processing. He received a Ph.D. in Computer Science from the Universi ty of Southampton in 2009.\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=192081 LOCATION:CERN 513-1-024 URL:http://indico.cern.ch/conferenceDisplay.py?confId=192081 END:VEVENT BEGIN:VEVENT SUMMARY:Nifty Native Implemented Functions: low-level meets high-level cod e DTSTART;VALUE=DATE-TIME:20120903T090000Z DTEND;VALUE=DATE-TIME:20120903T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-204810@cern.ch DESCRIPTION:\nErlang Native Implemented Functions (NIFs) allow developers to implement functions in C (or C++) rather than Erlang. NIFs are useful f or integrating high performance or legacy code in Erlang applications. The talk will cover how to implement NIFs\, use cases\, and common pitfalls w hen employing them. Further\, we will discuss how and why Erlang applicati ons\, such as Riak\, use NIFs.\n\n\n About the speaker\n\n Ian Plosker is the Technical Lead\, International Operations at Basho Technologies\, the makers of the open source database Riak. He has been developing software p rofessionally for 10 years and programming since childhood. Prior to worki ng at Basho\, he developed everything from CMS to bioinformatics platforms to corporate competitive intelligence management systems. At Basho\, he's been helping customers be incredibly successful using Riak.\n\nhttp://ind ico.cern.ch/conferenceDisplay.py?confId=204810 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=204810 END:VEVENT BEGIN:VEVENT SUMMARY:Wikimedia as a platform for scientific information DTSTART;VALUE=DATE-TIME:20120905T090000Z DTEND;VALUE=DATE-TIME:20120905T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-206721@cern.ch DESCRIPTION:\n During this presentation the topics that will touched upon include various ways in which wikis are being used in scientific research and publishing\, currently\, as well as some that are more speculative. Ma ny of the examples are drawn from the biological sciences\, and the talk i s intended to stimulate debate as to how the physics community - and CERN in particular - can enhance its interaction with the Wikimedia community\, or via Wikimedia with the public at large. For instance:\n\n PLOS Comput ational Biology Topic PagesWodak SJ\, Mietchen D\, Collings AM\, Russell R B\, Bourne PE (2012) "Topic Pages: PLoS Computational Biology Meets Wikipe dia". PLoS Comput Biol 8(3): e1002446\n \n Open Access Media Importer\n \ n A proposed Wiki Journal\, a peer-review journal to encourage academics to contribute Wikipedia articles\n \n Encyclopedia of Original Research a nd JATS-to-MediaWiki\n \n The Gene Wiki\n \n Wikigenes\n \n Wikis in Sc holarly Publishing\n \n The Journal of the Future\n\n His talk is being d rafted in public on Wikipedia\, see here.\n\n About the speaker\n\n Daniel Mietchen is a freelancing scientist\, trained in Biophysics at the Humbol dt University in Berlin. In addition to his conventional research\, he wor ks on the integration of research workflows with the World Wide Web\, part icularly by way of collaborative platforms like wikis. He is the Wikimedia n in Residence for Open Science\, hosted by the Open Knowledge Foundation\ , Germany.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=206721 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=206721 END:VEVENT BEGIN:VEVENT SUMMARY:Random Number Generation: A Practitioner's Overview DTSTART;VALUE=DATE-TIME:20120911T120000Z DTEND;VALUE=DATE-TIME:20120911T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-207559@cern.ch DESCRIPTION:\n We will look at random number generation from the point-of- view of Monte Carlo computations. Thus\, we will examine several serial me thods of pseudorandom number generation and two different parallelization techniques. Among the techniques discussed with be "parameterization\," wh ich forms the basis for the Scalable Parallel Random Number Generators (SP RNG) library. SPRNG was developed several years ago by the author\, and ha s become widely used within the international Monte Carlo community. SPRNG is briefly described\, and the lecture ends with a short revue of quasira ndom number generation. Quasirandom numbers offer many Monte Carlo applica tions the advantage of superior convergence rates. We also discuss plans f or a new version of SPRNG adapted to hybrid architectures.\nAbout the spea ker\n\nDr. Mascagni is full professor at Florida State University\, where he runs a research group consisting of post-doctoral associates\, graduate students\, and undergraduate workers. The areas they work on are parallel and distributed computing\, Grid computing\, random number generation\, M onte Carlo methods\, computational number theory and discrete algorithms\, and applications to materials science\, biochemistry\, electrostatics\, a nd finance.\n\n\nHe is on the editorial board of three journals in his fie ld\, and is a member of the ACM (Association of Computing Machinery)\, SIA M (Society of Applied Mathematics)\, and IMACS (International Association of Mathematics and Computers in Simulation). He is also a member of the Bo ard of Directors of IMACS. He has approximately 100 refereed technical pap ers that have appeared in a wide variety of publications in areas includin g Applied Mathematics\, Computer Science\, Simulation Science\, Monte Carl o Methods\, Computational Science\, High-Performance Computing\, Scientifi c Computing\, Computational Physics\, and Computational Neuroscience. He h as been a visiting professor at the University of Padova in Italy\, the Un iversity of Salzburg in Austria\, and the Swiss Federal Technical Institut e-Zürich in Switzerland\, and is a consultant to industry and government. He has made technical presentations in 18 countries and in most of the 50 U.S. states.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=207559 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=207559 END:VEVENT BEGIN:VEVENT SUMMARY:GPU acceleration of scientific applications: An update DTSTART;VALUE=DATE-TIME:20120924T090000Z DTEND;VALUE=DATE-TIME:20120924T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-206360@cern.ch DESCRIPTION:\nJust five years ago\, NVIDIA introduced CUDA\, the Compute U nified Device\nArchitecture\, which significantly simplified the developme nt of scientific\napplications on Graphics Processing Units (GPUs).\nSince these early days of CUDA\, both GPU hardware and software have made\ntrem endous progress in terms of programmability and ease of use\, allowing an\ never increasing set of applications to benefit from GPUs. \nEspecially t he latest generation of NVIDIA GPUs\, Kepler\, introduced many\nfeatures g eared towards HPC applications and to ensure high performance and\nease of use.\n\nIn this seminar\, we will give an overview of the Kepler architec ture\,\npresent the new features of CUDA 5 and show how they can be used t o\naccelerate scientific applications.\nIn addition\, we will present CARM A\, the CUDA on ARM development board\nhosting a NVIDIA Tegra CPU and CUDA GPU\, and show how to use these boards to\nexplore an exciting high-perfo rmance\, low power consumption node\narchitecture.\n\nAbout the speakers\n Edmondo Orlotti - HPC Development Manager\n\nEdmondo Orlotti is responsibl e for the business development of NVIDIA HPC\nSolutions in Southern Europe \, for Education and Research. He joined NVIDIA\nin 2005 as marketing mana ger for the same region\, then moving in 2008 to the\nprofessional solutio ns division. He's been addressing advanced visualization\nneeds of industr ial and academic environments\, integrating the portfolio of\nNVIDIA techn ologies with the local contexts of users needs.\n\nPeter Messmer - Senior Devtech Engineer\n\nPeter joined NVIDIA in 2011 after spending more than 1 5 years developing HPC\nand GPU accelerated applications for industry  an d Government clients\,\nmainly in the area of plasma and EM simulations\, data analysis and\nvisualization.  In his role as Senior Devtech Engineer at NVIDIA\, Peter is\nworking with HPC users around the globe supporting them in accelerating\ntheir scientific discovery process by taking\nadvant age of GPUs in their applications. Peter holds and MSc and PhD  in\nPhysi cs  from ETH Zurich\,\nSwitzerland\, with specialization in kinetic plasm a physics and non-linear\noptics.\n\nhttp://indico.cern.ch/conferenceDispl ay.py?confId=206360 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=206360 END:VEVENT BEGIN:VEVENT SUMMARY:SEJITS: embedded specializers to turn patterns-based designs into optimized parallel code DTSTART;VALUE=DATE-TIME:20120928T120000Z DTEND;VALUE=DATE-TIME:20120928T130000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-208891@cern.ch DESCRIPTION:\n All software should be parallel software. This is natural r esult of the transition to a many core world. For a small fraction of the world's programmers (efficiency programmers)\, this is not a problem. They enjoy mapping algorithms onto the details of a particular system and are well served by low level languages and OpenMP\, MPI\, or OpenCL. Most prog rammers\, however\, are "domain specialists" who write code. They are too busy working in their domain of choice (such as physics) to master the int ricacies of each computer they use. How do we make these programmers produ ctive without giving up performance?\n\n We have been working with a team at UC Berkeley's ParLab to address this problem. The key is a clear softwa re architecture expressed in terms of design patterns that exposes the con currency in a problem. The resulting code is written using a patterns-base d framework within a high level\, productivity language (such as Python). Then a separate system is used by a small group of efficiency programmers to translate that code into a highly optimized program. This software tran sformation system is called SEJITS: Selective Embedded Just In Time Specia lizers.\n\n In this talk\, I will describe the overall vision behind this work\, including the patterns that sit at its core. Most of the talk\, how ever\, will focus on SEJITS. We will explore how SEJITS works and show its use for a few applications. We'll close with some thoughts on where this technology might evolve and hopefully move into the mainstream of parallel computing.\n\n About the speaker\n\n Tim Mattson is a parallel programmer (Ph.D. Chemistry\, UCSC\, 1985). Tim has been with Intel since 1993 where he has worked with brilliant people on great projects such as:\n\n the f irst TFLOP computer (ASCI Red)\,\n \n the OpenMP API for shared memory pr ogramming\,\n \n the OpenCL programming language for heterogeneous platfo rms\,\n \n Intel's first TFLOP chip (the 80 core research chip)\, and\n \ n Intel's 48 core\, SCC research processor.\n\n Tim has published extensi vely including the books Patterns for Parallel Programming (with B. Sander s and B. Massingill\, Addison Wesley\, 2004)\, An Introduction to Concurre ncy in Programming Languages (with M. Sottile and C. Rasmussen\, CRC Press \, 2009)\, and the OpenCL Programming Guide (with A Munshi\, B. Gaster\, J . Fung\, and D. Ginsburg\, Addison Wesley\, 2011).\n\nhttp://indico.cern.c h/conferenceDisplay.py?confId=208891 LOCATION:CERN IT Auditorium URL:http://indico.cern.ch/conferenceDisplay.py?confId=208891 END:VEVENT BEGIN:VEVENT SUMMARY:The new Generation of Data Management DTSTART;VALUE=DATE-TIME:20130110T090000Z DTEND;VALUE=DATE-TIME:20130110T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-221366@cern.ch DESCRIPTION:\n Last year\, Software AG acquired Terracotta\, a leading pla yer for In Memory Data Management for the Enterprise. With it\, Software A G is bringing performance at any scale for the Business Application. Toget her with two other technologies\, CEP (Complex Event Processing) and Nirva na (a Low latency Messaging System)\, Software AG offers a visionary platf orm for Big Data and Cloud Computing.\n\n The presentation will focus on Terracotta and CEP\, addressing the following points:\n\n General present ation of Terracotta (Big Memory Max) & CEP Engine (Business Events)\n \n Explanation of the Terracotta architecture\, scalability and possible fiel ds of application\n \n CEP introduction\, integration with Terracotta\, a nalysis of real-time events\n\n For those interested in further details\, there will be an additional session in the afternoon (14-16h) about field s of application and interesting Use Cases in different industries and CER N\, namely:\n\n Examples of real use cases for Terracotta\n \n Code exam ples for Terracotta & CEP\n \n Possible fields of application for real-ti me measurement with CEP\n\n About the speakers\n\n Dr. Jürgen Krämer is a Vice President of Product Marketing at Software AG. He is responsible fo r all analytics and monitoring products. He has an extensive technology ba ckground and practical experience in data management and analytics. He pub lished various articles in this area\, is a speaker at international confe rences\, and a strategic advisor in business workshops. Dr. Jürgen Kräme r is a founder and managing director of RTM\, an award-winning university spinoff focused on real-time monitoring and since 2010 wholly owned subsid iary of Software AG. He holds a PhD with honors in Computer Science from U niversity of Marburg and received a scholarship from the German Federal Mi nistry of Economics and Technology and an executive education from the Eur opean Business School.\n\n Stephan Grotz is Solution Architect / Field Eng ineer for Terracotta DACH Region\, Terracotta is 100% owned by Software AG . Stephan's focus is Support to New Technologies Business as In-Memory Dat a Management with Terracotta\, Low-Latency Messaging with Nirvana and Comp lex Event processing (CEP)\n\nhttp://indico.cern.ch/conferenceDisplay.py?c onfId=221366 LOCATION:CERN BE Auditorium Meyrin URL:http://indico.cern.ch/conferenceDisplay.py?confId=221366 END:VEVENT BEGIN:VEVENT SUMMARY:CAPS OpenACC Compilers: Performance and Portability DTSTART;VALUE=DATE-TIME:20130227T100000Z DTEND;VALUE=DATE-TIME:20130227T110000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-233997@cern.ch DESCRIPTION:\n The announcement late 2011 of the new OpenACC directive-bas ed programming standard supported by CAPS\, CRAY and PGI compilers has ope n up the door to more scientific applications that can be ported on many-c ore systems. Following a porting methodology\, this talk will first review the principles of programming with OpenACC and then the advanced features available in the CAPS compilers to further optimize OpenACC applications: library integration\, tuning directives with auto-tune mechanisms to buil d applications adaptive to different GPUs. CAPS compilers use hardware ven dors' backends such as NVIDIA CUDA and OpenCL making them the only OpenACC compilers supporting various many-core architectures.\n\n About the speak er\n\n Stéphane Bihan is co-funder and currently Director of Sales and Ma rketing at CAPS enterprise. He has held several R&D positions in companie s such as ARC international plc in London\, Canon Research Center France\, ACE compiler experts in Amsterdam and the INRIA research lab. His work wa s devoted to the development of software development tools for specialized processors.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=233997 LOCATION:CERN TH Conference Room URL:http://indico.cern.ch/conferenceDisplay.py?confId=233997 END:VEVENT BEGIN:VEVENT SUMMARY:Crushing data silos with ownCloud DTSTART;VALUE=DATE-TIME:20130314T083000Z DTEND;VALUE=DATE-TIME:20130314T093000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-233993@cern.ch DESCRIPTION:\n More and more people store their personal files and documen ts in cloud services like Dropbox\, Google Drive\, Skydrive or iCloud. The reason is that they provide convenient features to sync your files betwee n devices and share them with others. We are heading full speed into a fut ure where a huge piece of the personal information of the world is stored in very few centralized services. Questions emerge what the impact on user privacy\, surveillance\, lawfulness of content and storage cost will be i n in the long run. I don't think that a world where most of the personal d ata of the world is stored on servers of a hand full companies is a good o ne.\n\n This talk will discuss the problems of a future with centralized c loud file sync and share services and will present ownCloud as a possible solution.\n\n ownCloud is a free software project that offers a decentrali zed alternative to proprietary cloud services where everybody can run an o wn cloud service comparable with Dropbox but on own hardware and with full control. ownCloud provides all the features of the proprietary services l ike syncing to Mac\, Windows Linux\, iOS and Android\, sharing files with others\, photo gallery\, media player and more. But ownCloud is free softw are\, available at no cost\, can provide all the state of the art cloud fe atures\, full privacy and run on personal hardware.\n\n About the speaker\ n\nFrank Karlitschek is a long time open source contributor and former boa rd\nmember of the KDE e.V. He managed engineering teams for over 10 years in\ndifferent companies. In 2010 he started the ownCloud project and is le ading\nthe community project since then. In 2011 he founded ownCloud Inc. together\nwith Markus Rex and Holger Dyroff from SUSE to offer commercial services\,\nfor enterprises around ownCloud.\n\n\nhttp://indico.cern.ch/co nferenceDisplay.py?confId=233993 LOCATION:CERN TH Conference Room URL:http://indico.cern.ch/conferenceDisplay.py?confId=233993 END:VEVENT BEGIN:VEVENT SUMMARY:Application security: Not so obvious vulnerabilities DTSTART;VALUE=DATE-TIME:20130321T153000Z DTEND;VALUE=DATE-TIME:20130321T163000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-241705@cern.ch DESCRIPTION:\n This talk will present several little known but interesting situations where security vulnerabilities may arise. Basic XSS and SQL in jection vulnerabilities will not be covered\, but the bypass of strict def ensive filters (either server-side or via a WAF) will be studied. The weir d behaviors of PHP when using lax operators ("==" instead of "===") will b e shown. We'll also get into XML technologies like Web Services\, digital signatures\, SVG documents and the like\, covering attacks like denial of service\, XML External Entities\, Server Side Request Forgery or remote co de execution via XSLT. Regarding fuzzing\, one of my recent campaigns will be discussed. Most of the examples and demonstrations are based on real-w orld code. I promise a lot of acronyms\, some really nifty bugs and maybe a few good laughs.\n\n About the speaker\n\n Nicolas Gregoire has more tha n 12 years of experience in penetration testing and auditing of networks a nd (mostly Web) applications. He founded Agarri\, a small company where he finds security bugs for customers and for fun. His research was presented at numerous conferences around the world (Hack in the Box\, HackInParis\, ZeroNights\, ...) and he was publicly thanked by some well known vendors (Microsoft\, Adobe\, Mozilla\, Google\, Apple\, VMware\, ...) for responsi bly disclosing vulnerabilities in their products.\n\nhttp://indico.cern.ch /conferenceDisplay.py?confId=241705 LOCATION:CERN Salle conference BE URL:http://indico.cern.ch/conferenceDisplay.py?confId=241705 END:VEVENT BEGIN:VEVENT SUMMARY:Discussion on the Future of C++ (Bristol Meeting) DTSTART;VALUE=DATE-TIME:20130404T090000Z DTEND;VALUE=DATE-TIME:20130404T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-240073@cern.ch DESCRIPTION:\n The future of C++ is defined by an ISO committee that inclu des CERN. Several proposals are on the table for the upcoming committee me eting (Bristol\, April 15-20). Axel will present how the standardization p rocess works\, and show the main proposals for upcoming standard revisions (the next one being scheduled already for next year!) and technical speci fications (the first one is already expected this year!). The complete lis t of proposals is attached to this event page. The feedback from this meet ing will then be voiced at the standards meeting: participate here and you help shape C++!\n\n About the speaker\n\n Axel Naumann is a member of the ROOT team in the PH-SFT group\, where he is working on ROOT's interpreter and type information system. Since 2012 CERN sends him to the C++ standar d meetings.\n\nhttp://indico.cern.ch/conferenceDisplay.py?confId=240073 LOCATION:CERN 32-1-A24 URL:http://indico.cern.ch/conferenceDisplay.py?confId=240073 END:VEVENT BEGIN:VEVENT SUMMARY:Big Process for Big Data DTSTART;VALUE=DATE-TIME:20130502T090000Z DTEND;VALUE=DATE-TIME:20130502T100000Z DTSTAMP;VALUE=DATE-TIME:20130522T124621Z UID:indico-event-239385@cern.ch DESCRIPTION:\nLarge and diverse data result in challenging data management problems that\nresearchers and facilities are often ill-equipped to handl e. I propose a new\napproach to these problems based on the outsourcing of research data\nmanagement tasks to software-as-a-service providers. I arg ue that this\napproach can both achieve significant economies of scale and accelerate\ndiscovery by allowing researchers to focus on research rather than mundane\ninformation technology tasks. I present early results with the approach in\nthe context of Globus Online.\n\n\nhttp://indico.cern.ch/ conferenceDisplay.py?confId=239385 LOCATION:CERN Council Chamber URL:http://indico.cern.ch/conferenceDisplay.py?confId=239385 END:VEVENT END:VCALENDAR