Computing and Astroparticle Physics - 2nd ASPERA Workshop - hosted by Port d'Informació Científica (PIC)

30 May 2011, 08:30 → 31 May 2011, 14:10 Europe/Madrid

Aules 11-13 (Casa Convalescència, Barcelona)

Manuel Delfino (Port d'Informació Científica (PIC) - Universitat Autònoma de Barcelona (UAB)), Stavros Katsanevas

The second ASPERA Workshop on Computing and Astroparticle Physics will take place in Barcelona, Spain organized by the Port d'Informació Científica (PIC). It will focus on some of the issues identified in the first workshop in Lyon.
The workshop is conceptualized as a dialog between Models for Computing and Data Pipelines for astroparticle projects and Technologies for Data Processing and Computing. Future workshops will be planned to focus on other issues. We call the "actors" of this second workshop the "Modellers" and the "Technologists". A novel feature we will try out in this workshop is to have Parallel sessions on Models with rapporteurs who are Technologists, and viceversa, sessions on Technologies with rapporteurs who are Modellers. See the programme for all the details.
The workshop is by invitation, prospective participants can apply via the Registration section of the Indico page. The workshop will have invited summary and rapporteur talks, as well as short presentations in the parallel sessions which prospective participants can propose via the "Call for Abstracts" section of the Indico page.
NOTE ADDED: Unfortunately the ASPERA Indico requires a CERN password to submit an abstract. If you don't have one, please send the abstract by e-mail to the conference secretary Ada Trilla-Kessler atrilla@pic.es
The workshop itself is free of charge (costs covered by ASPERA), participants will have to cover their travel and hotel expenses. UPDATE: The Workshop Dinner on Monday night will be sponsored by ASPERA and thus free of charge to participants.
The workshop co-organizers are Stavros Katsanevas (ASPERA/IN2P3) and Manuel Delfino (PIC/UAB). For any further information, please contact the co-organizers via the Workshop Secretary, Ms. Ada Trilla-Kessler e-mail atrilla@pic.es .

Monday, 30 May

08:30 Registration

Welcome and Introduction

1 Welcome and Introduction

Speakers: Prof. Manuel Delfino Reznicek (Port d'Informació Científica (PIC)), Stavros KATSANEVAS (CNRS/IN2P3)

Slides Aspera_delfino_welcome.pdf

Models for Computing and Data Pipelines (plenary)

2 The Basic Elements of a Computing and Data Processing Model: an LHC perspective

Speaker: Dr Gonzalo Merino (Port d'Informació Científica (PIC))

Slides 20110530_ASPERA_GM.pdf

3 Overview of Models for Event Experiments (Cosmic and Gamma Rays, Neutrinos and Dark Matter)

Speaker: Dr Giovanni Lamanna (LAPP, Université de Savoie et IN2P3/CNRS)

Slides ASPERA_Barcelona_Lamanna.pdf

4 Overview of Models for Surveys and Dark Energy Experiments

Speaker: Pierre Astier (LPNHE/CNRS/University of Paris)

Slides wide-field-data-reduction.pdf

5 Overview of Models for Gravitational Wave Experiments

I'll discuss the large-scale data analysis methods and hardware currently being used for gravitational wave detection projects, and being planned for 2nd and 3rd generation detectors. I'll also talk about technology that would help to improve these efforts.

Speaker: Prof. Bruce Allen (Max Planck Society/Albert Einstein Institute Hannover)

Slides Aspera_Barcelona_2011.key

11:30 Coffee Break

Technologies for Data Processing and Computing (plenary)

6 Overview of Computing Technologies (Virtualization, Cloud, Grid)

Speaker: Dr Pere Mato (CERN)

Slides Virtualization&Clouds.pdf Virtualization&Clouds.pptx

7 Overview of Database Technologies

Speaker: Eric Grancher (CERN)

Slides 20110530_ASPERA-EG-09.pdf 20110530_ASPERA-EG-09.pptx

8 Overview of Data Storage, Access and Distribution Technologies

Speaker: Dr Dirk Duellmann (CERN)

Slides dirkd-aspera-storage.pdf

13:30 Standing-up Lunch Break

Models in Event-driven Experiments

9 Computing for the Fermi Large Area Telescope

Speaker: Dr Johann Cohen-Tanugi (LUPM, Université Montpellier2/CNRS-IN2P3)

Slides LATComputingModel_ASPERA_30mai2011.pdf

10 EE Discussion 1

11 The IceCube Data Pipeline

Speaker: Dr Martin Merck (University of Wisconsin-Madison)

Slides

• 1st Event
• 2nd Event
• Merck_ASPERA_Workshop.pdf

12 EE Discussion 3

13 Models for open observatories and the multi-PB scale

I will briefly review the computing model that have been used for high-energy (and other) open observatories and the technologies that could be used to face the access to multi-PB scale storage.

Speaker: Dr Roland Walter (ISDC, Observatoire de Geneve)

Slides Barcelona.pdf

14 EE Discussion 2

Models in Surveys and Dark Energy experiments

15 Computing model for Large Cosmological Surveys and Simulations: PAU and MICE

We present a computing model designed for two cosmological projects, the PAU survey and the MICE simulations. The Physics of the Accelerating Universe (PAU) project aims at the study of the acceleration in the expansion of the universe. The PAU Team is building an instrument called PAUCam designed to be located at the William Herschel Telescope in La Palma. The camera will produce 200GB of Raw data per observing night for a total of 50TB and it will be transferred and stored at PIC in a nightly basis through the RedIRIS network. The PAU data management group is developing a specific reduction software for the PAU data. The software is organized into different pipelines: pixel simulation, nightly reduction, quality analysis and multi-epoch. The PAUdm software uses a combination of public (Astromatic) and self packages implemented in python. It has been designed in order to work both under a GRID environment and in a local PC (for community users) and to be flexible to accept data from external surveys. Parallelization is required due to time constraint allowing simultaneous image reduction. A PostgreSQL data base supports the control of all raw and processed data, providing access to the production catalogues for the PAU community. The MICE (http://www.ice.cat/mice) cosmological simulations, produced at the Marenostrum Supercomputer of the Barcelona Supercomputer Center (BSC), have been transferred and stored at PIC. They serve as input for the pixel simulation, for the scientific error estimation and to optimize the survey strategy, the calibration and reduction pipelines and the photo-z analysis of the galaxy surveys in which we actively participate (PAU,DES,Euclid).

Speaker: Dr Nadia Tonello (Port d'Informació Científica (PIC))

Slides PAU_MICE_aspera_workshop.pdf

16 DE Discussion 1

17 Computing model for the Spanish contribution to the EUCLID Simulation OU

Speaker: Dr Santiago Serrano Elorduy (Instituto de Ciencias del Espacio (ICE-IEEC-CSIC))

Slides EUCLID_aspera_workshop.pdf

18 DE Discussion 2

19 Current experience with the SNLS data model

Speaker: Pierre Astier (LPNHE/CNRS/University of Paris)

Slides snls-data-reduction.pdf

20 DE Discussion 3

Models in Gravitational Wave Experiments

21 LIGO Global Computing in the next decade

Speaker: Prof. Patrick Brady (University of Wisconsin-Milwaukee)

Slides LIGO-Data-Grid-Aspera-2011.pdf

22 GW Discuss 1

23 The VIRGO Computing Model: present status and future plans

Speaker: Gergely Debreczeni (KFKI Research Institute for Particle and Nuclear Physics)

Slides Virgo-Debreczeni.pdf

24 GW Discuss 2

25 The computing model for the Einstein Telescope

Speaker: Prof. Bangalore Sathyaprakash (Cardiff University)

Slides ET-ComputationalChallenges-110529.pdf

26 GW Discuss 3

17:30 Coffee Break

Technologies for Data Processing and Computing (plenary)

27 Graphs: in the eye of the storm of knowledge management

1. Graph technology and management. 2. Small, large and huge graphs. A perspective of the evolution of graph needs. 3. Scientific and knowledge data, management and integration of sources. How the graph can help. 4. Applications on the management of large and huge graphs. 5. Conclusions.

Speaker: Prof. Josep Lluís Larriba Pey (Data Management Group, Universitat Politècnica de Catalunya)

28 Questions

Technologies for Computing

29 CernVM – a virtual software appliance for LHC applications

Speaker: Dr Pere Mato (CERN)

Slides CernVM.pdf CernVM.pptx

30 COMPU Discuss 1

31 WNoDeS, a flexible and scalable Grid/Cloud virtualization system

CNAF is the national computing facility of INFN. It acts as the Italian Tier-1 for the LHC experiments and as a key computing center for several other experiments as well, including astroparticle physics ones such as MAGIC, PAMELA, AMS, Argo, Auger and Virgo. Many of these experiments call for more flexibility in their execution environment than is normally provided by traditional resource centers. For example, it is sometimes desirable to execute jobs on compute nodes especially tailored to the experiment needs. In this presentation we will outline the general characteristics of the Grid/Cloud virtualization system (called WNoDeS) we have developed within INFN and deployed at CNAF, where it has been running in production since November 2009. WNoDeS currently handles about 2000 Virtual Machines out of the CNAF global resource pool, and is heavily used by several astroparticle experiments. We will show, in particular, the setup (and its possible evolutions) used by the Auger experiment to access a condition database through custom Virtual Machines managed by WNoDeS.

Speaker: Davide Salomoni (INFN-CNAF)

Slides salomoni_wnodes_110530.pdf

32 COMPU Discuss 2

Technologies for Data Storage

33 dCache with Enstore: A cost-effective system for Petabyte data storage

Speaker: Francisco Martínez (Port d'Informació Científica (PIC))

Slides 20110531_dCache_for_ASPERA.odp 20110531_dCache_for_ASPERA.pdf

34 STOR Discuss 1

35 GEMSS, the Grid Enabled Mass Storage System

CNAF is the central computing facility of INFN. In particular it is the Italian Tier-1 for the experiments at LHC and the main Italian computing center also for several other experiments, notably including astro and space physics ones such as MAGIC, PAMELA, AMS, ARGO, AUGER and VIRGO. One of the most critical issues for a computing centre dealing with several PB of data is the management of the Mass Storage System (MSS). A MSS called GEMSS, based on the interaction between GPFS and TSM by IBM, has been developed at CNAF, and it is now in production since some years for all the experiments. In this talk we aim to show the excellent performance, stability and ease of management of GEMSS.

Speaker: Luca Dellagnello (CNAF-Istituto Nazionale Fisica Nucleare (INFN)-Unknown)

Slides lda-aspera-20110530.pdf

36 STOR Discuss 2

37 A data transfer system for MAGIC based on gLite FTS and multiVAC

Speaker: Roger Firpo Curcoll (Port d'Informació Científica)

Slides RFirpo_DataTransfer_ASPERA.pdf

38 STOR Discuss 3

Technologies for Pipeline Automation and Database Access

39 The KATRIN Analysis Framework

The KArlsruhe TRItium Neutrino experiment (KATRIN) will investigate the spectral shape of beta-decay electrons close to their kinematic endpoint in order to determine the neutrino rest mass in a model-independent method with an unprecedented sensitivity of 0.2 eV/c^2. For this purpose KATRIN makes use of a molecular gaseous tritium source, a transport section with differential and cryogenic pumps, two electrostatic spectrometers with magnetic adiabatic collimation (MAC-E filter) and a segmented silicon detector. Apart from detector DAQ a large number of heterogeneously distributed sensors has to be constantly monitored by a mixed PSC7 and Fieldpoint / Labview based slow-control system. The system has to maintain highly stable operating conditions over several years in order to reach the challenging sensitivity goal. As the experiment approaches the commissioning phase, the KATRIN analysis framework is currently under development. The server-side data processing will be handled by a central C++ web service, employing an object-relational mapping (ORM) database layer, XML data binding and an intuitive remote procedure call (RPC) interface to the actual client-side, ROOT based analysis applications.

Speaker: Mr Marco Haag (University of Karlsruhe, EKP)

Slides AnalysisFramework_Barcelona.pdf AnalysisFramework_Barcelona.pptx

40 DB Discuss 1

41 Automating Data Pipelines with DIRAC: from LHCb to Astroparticles

Speaker: Ricardo Graciani Diaz (ICC, Universtat de Barcelona)

Slides DIRAC-Data-Processing.pdf

42 DB Discuss 2

43 Management of Simulation Production in Auger using Grid Technology

Speaker: Dr Julio Lozano Bahilo (Universidad de Granada)

Slides AugerProductionGrid1.pdf

44 WARP Database Organization

Speaker: Dr Natalia Deniskina (University of federico II, Naples)

Slides warp_database_deniskina.pdf

20:00 Free Time

21:00 Workshop Dinner sponsored by ASPERA

Tuesday, 31 May

Model challenges and opportunities seen by Technologists

45 Technologist's Round-Table and Discussion: Model Challenges

Speaker: Dr Pere Mato (CERN)

Technology opportunities seen by Modellers

46 Modeller's Round-Table and Discussion: Technology Opportunities

Speaker: Bruce Allen (Albert Einstein Institute)

Slides Aspera_Barcelona__second_day_2011.pdf

12:00 Coffee Break

Summary and Discussion

47 Hot topics and hot issues lists: Discussion

Speaker: Prof. Manuel Delfino (Port d'Informació Científica (PIC))

Slides Aspera_delfino_Discussion.pdf

48 Summary of Summaries and Topics for Next Workshop

Speaker: Stavros KATSANEVAS (CNRS/IN2P3)

Slides 2011may_barcelona_computing_SK.pdf 2011may_barcelona_computing_SK.pptx