{\rtf1\mac\ansicpg10000\cocoartf824\cocoasubrtf230
{\fonttbl\f0\fnil\fcharset77 LucidaGrande-Bold;\f1\fnil\fcharset77 LucidaGrande;\f2\fnil\fcharset77 Verdana-Italic;
\f3\fnil\fcharset77 Verdana-BoldItalic;}
{\colortbl;\red255\green255\blue255;\red214\green78\blue46;\red143\green24\blue13;\red255\green0\blue0;
\red159\green156\blue114;\red32\green95\blue186;\red133\green144\blue160;}
{\*\listtable{\list\listtemplateid1\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid1}
{\list\listtemplateid2\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid2}
{\list\listtemplateid3\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid3}
{\list\listtemplateid4\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid4}
{\list\listtemplateid5\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid5}
{\list\listtemplateid6\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid6}
{\list\listtemplateid7\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid7}
{\list\listtemplateid8\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid8}
{\list\listtemplateid9\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid9}
{\list\listtemplateid10\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid10}
{\list\listtemplateid11\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid11}
{\list\listtemplateid12\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid12}
{\list\listtemplateid13\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid13}
{\list\listtemplateid14\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid14}
{\list\listtemplateid15\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid15}
{\list\listtemplateid16\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid16}
{\list\listtemplateid17\listhybrid{\listlevel\levelnfc0\levelnfcn0\leveljc2\leveljcn2\levelfollow0\levelstartat1\levelspace360\levelindent0{\*\levelmarker \{decimal\}.}{\leveltext\leveltemplateid0\'02\'05.;}{\levelnumbers\'01;}}{\listname ;}\listid17}}
{\*\listoverridetable{\listoverride\listid1\listoverridecount0\ls1}{\listoverride\listid2\listoverridecount0\ls2}{\listoverride\listid3\listoverridecount0\ls3}{\listoverride\listid4\listoverridecount0\ls4}{\listoverride\listid5\listoverridecount0\ls5}{\listoverride\listid6\listoverridecount0\ls6}{\listoverride\listid7\listoverridecount0\ls7}{\listoverride\listid8\listoverridecount0\ls8}{\listoverride\listid9\listoverridecount0\ls9}{\listoverride\listid10\listoverridecount0\ls10}{\listoverride\listid11\listoverridecount0\ls11}{\listoverride\listid12\listoverridecount0\ls12}{\listoverride\listid13\listoverridecount0\ls13}{\listoverride\listid14\listoverridecount0\ls14}{\listoverride\listid15\listoverridecount0\ls15}{\listoverride\listid16\listoverridecount0\ls16}{\listoverride\listid17\listoverridecount0\ls17}}
\margl1440\margr1440\vieww19040\viewh15020\viewkind0
\deftab720
\pard\pardeftab720\sl520\sa40\ql\qnatural

\f0\b\fs50\fsmilli25200 \cf2 Baseline Services workgroup December 02 \
\
\pard\pardeftab720\sl260\sa20\ql\qnatural

\fs26\fsmilli13200 \cf3 Present:
\fs50\fsmilli25200 \cf2 \
\pard\pardeftab720\sl360\ql\qnatural

\f1\b0\fs24 \cf0 Alice\
Dirk D.\
LHCb \
Erwin\
ARDA (Julia) \
ATLAS \
Steve Traylen\
Flavia\
Markus\
Ian 
\f0\b\fs50\fsmilli25200 \cf2 \

\f1\b0\fs24 \cf0 CMS (Stephano Belforte)
\f0\b\fs50\fsmilli25200 \cf2  \
\pard\pardeftab720\sl520\sa40\ql\qnatural
\cf2 \
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls1\ilvl0
\fs24 \cf0 The session started with a lively discussion on the relation between the Baseline Services Work Group and the TCG.\
\
Ian:\
I would like how we can stop and wrap up. \
Flavia's summary is a good point to start.\
We should go through this and decide what we want to write up.\
Maybe we start with this first.\
\
LHCB:\
The list is very useful, but we should not base our report on this.\
\
The group went through Flavia's document:\
Comments are indicated by -------> and the text color is \cf4 red\cf0 \
\pard\pardeftab720\sl520\sa40\ql\qnatural

\fs50\fsmilli25200 \cf2 -------------------------------------\
Summary of Open Issues reported by LHC experiments\
\pard\pardeftab720\sl360\ql\qnatural

\f1\b0\fs24 \cf0 \
\pard\pardeftab720\sl360\sa120\ql\qnatural
\cf5 \cb5 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \cb1 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 1. Security, authorization, authentication\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls2\ilvl0
\fs24 \cf0 {\listtext	1.	}VOMS available and stable     \
{\listtext	3.	}VOMS groups and roles used by all middleware
\f1\b0  \uc0\u8232 
\f2\i Support for up to o(10) groups.
\f1\i0 \
\ls2\ilvl0
\f0\b {\listtext	4.	}VOMS supporting user metadata -------------> comes from LHCb : Nickname to each user to identify users home directories. All VO/user management info in one place.  Has to be understood what this means for the service (load) 
\f1\b0  \uc0\u8232 
\f2\i Storing arbitrary user metadata should be possible in VOMS with an easy
\f1\i0  \uc0\u8232 
\f2\i interface to access the user parameters, e.g. passing them in the VOMS proxy
\f1\i0  \
\ls2\ilvl0
\f0\b {\listtext	5.	}Automatic handling of service proxy renewal
\f1\b0  \uc0\u8232 
\f2\i The user should not need to know which server to use to register
\f1\i0  \uc0\u8232 
\f2\i his proxy for a specific service.
\f1\i0  \
\ls2\ilvl0
\f0\b {\listtext	6.	}Recommendations on how to develop experiment specific secure services
\f1\b0  \uc0\u8232 
\f2\i Best framework to write a secure service interacting with the Grid
\f1\i0  \uc0\u8232 
\f2\i using delegated and automatically renewed user credentials;
\f1\i0  \uc0\u8232 
\f2\i API or "development guide" for security delegation standards and
\f1\i0  \uc0\u8232 
\f2\i documentation;
\f1\i0  \uc0\u8232 
\f2\i GSI delegation vs. Myproxy, GT2 vs. GT4 vs. Web services, etc.
\f1\i0                             \cf4  -----------> Document exists, but tool kit needed\cf0 \
\ls2\ilvl0
\f0\b {\listtext	7.	}Service needed for automatic renewal of Kerberos credentials via the Grid        \cf4  --------> Alice:  automatic renewal 
\f1\b0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 2. Information System\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls3\ilvl0
\fs24 \cf0 {\listtext	1.	}Stable access to static information
\f1\b0                                                         \cf4  ---------------> split between static and dynamic information\cf0 .           \uc0\u8232 
\f2\i Grid Information System (BDII or equivalent) should provide a stable
\f1\i0      \cf4 ---------------> caching might be a solution \uc0\u8232 
\f2\i \cf0 access to the static information (services end-points and characteristics).
\f1\i0  \cf4  -------------> glue as in gLite  \cf0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 3. Storage Management\
\pard\pardeftab720\sl360\ql\qnatural

\f1\b0\fs24 \cf0 \
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls4\ilvl0
\f0\b \cf0 {\listtext	1.	}SRM interface provided by all Storage Element Services
\f1\b0                        \uc0\u8232 
\f2\i SRM must be a fully supported specification as indicated in
\f1\i0  \uc0\u8232 {\field{\*\fldinst{HYPERLINK "http://lcg.web.cern.ch/LCG/peb/BS/BSReport-v1.0.pdf"}}{\fldrslt \cf6 \ul \ulc6 Baseline Service group report}}. \uc0\u8232 
\f2\i In particular, the functionalities provided with SRM V2.1.1 are requested.
\f1\i0  \uc0\u8232 
\f2\i Mostly needed are: space reservation, file pinning, bulk operations.
\f1\i0  \
\ls4\ilvl0
\f0\b {\listtext	2.	}Common and homogeneous functionality (same semantic) for all Storage Services
\f1\b0  \uc0\u8232 
\f2\i Smooth transition from SRM v1 to SRM v2.
\f1\i0  \cf4  -----------------> different APIs . Test is needed to ensure that the SRM is complient.\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls4\ilvl0\cf0 {\listtext	3.	}\cf4 different problem: transition.  LHCb: V2 is fine. ALICE: There was something that went away. Information on protocol is missing\cf0 . \
{\listtext	4.	}Ian: V1 and V2 have to be maintained in parallel.  LHCb: gFal level  ATLAS: FTS would be a problem.  \uc0\u8232 
\f2\i SE interoperabiliy issues must be solved.
\f1\i0                                  \cf4      ----------> Functionality must be homogenous.\uc0\u8232 
\f2\i \cf0 Applications must be able to access SRM functionalities at sites.    \cf4 ------------->  SRM client lib. 
\f1\i0 \
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls4\ilvl0
\f0\b \cf0 {\listtext	5.	}Support for quota management
\f1\b0  \uc0\u8232 
\f2\i Support for quota management both at group and user level should be offered
\f1\i0  \cf4   -------------> yes \uc0\u8232 
\f2\i \cf0 by all Storage Services (requested in particular by
\f1\i0  
\f3\i\b ATLAS
\f1\i0\b0  
\f2\i ,
\f1\i0  
\f3\i\b CMS
\f1\i0\b0  \uc0\u8232 
\f2\i and
\f1\i0  
\f3\i\b LHCB
\f1\i0\b0  
\f2\i ).
\f1\i0  \
\ls4\ilvl0
\f0\b {\listtext	6.	}Checking of the file validity after the new replica creation.   \cf4  ----------------> checksum: each copy has to perform a checksum  (on demand) (minimum file lentgth)\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls4\ilvl0
\f1\b0 \cf4 LHCb and ATLAS: Remove and other operations have to be validated that they have had correct effect on the fabric \cf0 \
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls4\ilvl0
\f0\b \cf0 {\listtext	8.	}Highly optimized SRM client tools                    \cf4   --------> LHCb: no direct access to the BDII should be required for any operation 
\f1\b0  \uc0\u8232 
\f2\i \cf0 SRM clients should be based on a highly optimized C/C++ library (gfal).
\f1\i0      \uc0\u8232 
\f2\i In particular, command line tools based on the C/C++ API (and not java based)
\f1\i0  \uc0\u8232 
\f2\i should be available. Python binding is required.
\f1\i0     \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 4. Data Management\
\pard\pardeftab720\sl360\ql\qnatural

\f1\b0\fs24 \cf0 \
\pard\pardeftab720\sl220\sa20\ql\qnatural

\f0\b\fs22\fsmilli11400 \cf3 4.1 File Transfer Service\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls5\ilvl0
\fs24 \cf0 {\listtext	1.	}File Transfer Service
\f1\b0  \uc0\u8232 
\f2\i Available on all SC3 sites on WNs and VOBOXes --\cf4 ------------> Ian: Clients????? Flavia: Only clients!
\f1\i0 \cf0 \
\ls5\ilvl0
\f0\b {\listtext	2.	}Reliability
\f1\b0  \uc0\u8232 
\f2\i Keep retrying until told to stop. Allow for real-time monitoring of
\f1\i0  \uc0\u8232 
\f2\i errors for transfer (parseable errors preferable) so that reshuffling of
\f1\i0  \uc0\u8232 
\f2\i transfers, cancellation, etc. is possible.
\f1\i0  \uc0\u8232 
\f2\i Signal conditions such as source missing, destination down, etc.
\f1\i0  \
\ls5\ilvl0
\f0\b {\listtext	3.	}A service is needed for automatic file transfers betwen two sites on the Grid
\f1\b0  \uc0\u8232 
\f2\i Start the transfers giving as input information the name of the SE
\f1\i0  
\f2\i (source and destination) and the file SURL (note: the file transfer service
\f1\i0  
\f2\i should not be linked to any specific catalogue; the SURL is the best
\f1\i0  
\f2\i specification for the file)
\f1\i0 \
\ls5\ilvl0
\f0\b {\listtext	4.	}Central service for all transfers
\f1\b0  \uc0\u8232 
\f2\i FTS is needed to provide a single central entry point to all the required
\f1\i0  \uc0\u8232 
\f2\i transfer channels including T0-T1, T1-T1 and T1-T2/T2-T1 transfers and for the T2
\f1\i0  \uc0\u8232 
\f2\i sites running analysis tasks.
\f1\i0  \
\ls5\ilvl0
\f0\b {\listtext	5.	}FTS should handle the automatic proxy renewal if necessary
\f1\b0 \
\ls5\ilvl0
\f0\b {\listtext	6.	}FTS "improvements" and feature requests as specified in the FTS workshop
\f1\b0  \uc0\u8232 
\f2\i Please, check:
\f1\i0  \uc0\u8232 {\field{\*\fldinst{HYPERLINK "http://agenda.cern.ch/fullAgenda.php?ida=a056842"}}{\fldrslt \cf6 \ul \ulc6 FTS Workshop agenda and minutes}}\
\ls5\ilvl0
\f0\b {\listtext	7.	}SRM interface fully integrated within FTS
\f1\b0  \uc0\u8232 
\f2\i Possibility to specify type of space, lifetime of a pinned file, etc.
\f1\i0  \
\ls5\ilvl0
\f0\b {\listtext	8.	}Support priorities, with possibility to do late reshuffling
\f1\b0 \
\ls5\ilvl0
\f0\b {\listtext	9.	}Support for plug-ins to allow interactions with experiment's services
\f1\b0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl220\sa20\ql\qnatural

\f0\b\fs22\fsmilli11400 \cf3 4.2 File Transfer Service\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls6\ilvl0
\fs24 \cf0 {\listtext	1.	}FPS plug-ins for VO specific agents
\f1\b0  \uc0\u8232 
\f2\i FPS should provide easy plug-in of the VO specific agents to implement retry
\f1\i0  \uc0\u8232 
\f2\i policies in case of any kind of failure.
\f1\i0  \
\ls6\ilvl0
\f0\b {\listtext	2.	}FPS should handle higher level operations
\f1\b0  \uc0\u8232 
\f2\i FPS should handle higher level operations such as data routing if necessary;
\f1\i0  \uc0\u8232 
\f2\i replication operations (without specification for the file source);
\f1\i0  \uc0\u8232 
\f2\i File Transfer Requests with multiple destination sites.
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf4 --------------->Ian :   Leave this for the workshop diskussion \
\pard\pardeftab720\sl220\sa20\ql\qnatural

\f0\b\fs22\fsmilli11400 \cf3 4.3 Grid File Catalogue Service\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls7\ilvl0
\fs24 \cf0 {\listtext	1.	}LFC as global and local file catalogue
\f1\b0  \uc0\u8232 
\f3\i\b CMS
\f1\i0\b0  
\f2\i is using LFC as global file catalogue for current MC production (phased out during 2006).
\f1\i0  \uc0\u8232 
\f2\i Expected access rate: 100Hz peak, few Hz average as file lookup.
\f1\i0  \
\ls7\ilvl0
\f0\b {\listtext	2.	}LFC requested features
\f1\b0  
\f2\i Support for replica attributes: tape, tape wth cache, pinned cache, disk,
\f1\i0  \uc0\u8232 
\f2\i archived tape, etc.
\f1\i0  \uc0\u8232 
\f2\i Custodial flag.
\f1\i0    \cf4 _____------>  The concept of Master Copy that can't be deleted. Ian: Limited, static bit of metadata for each replica. No schema evolution (for replicas)    LHCb: keep it lean  CMS: If this is implemented we will use it, if not we will not use it ever!!! Site has to update this. Ian: No the experiments own the access. CMS: For steering the job to the place where the data is on disk Ian: This is a problem. There is no update of the replica catalouge with the state of the file at the site. CMS: 99% of the time this is no problem.  Ian: This information you can get from the SRM. CMS: Yes Ian: Is this then still useful? Alice: Yes. There is a lot of static information. LHCb: Keep the catalogue as lightweight as possible.  Ian: Ask JeanPhillipe what the costs are.\cf0 \uc0\u8232 
\f2\i Dump of catalog entries: e.g. dump of all actions, new registered files, etc. 
\f1\i0    \cf4   ---> Atlas: not sure we need this anymore, was more of a checking tool. FTS can provide the information.\uc0\u8232 
\f0\b \cf0 	4.	POOL interface to LFC  \cf4  ---> Atlas: Still an issue (filesize is not exposed by pool) Dirk: we can only expose the common set, we can expose technology based information. We can only give you what a substantial subset of the catalogues expose, or abstraction becomes important. Atlas now talks to LFC and through pool to LFC.  Dirk: The catalogue abstraction should NOT be in pool, but in the middleware.  Markus: We need an SRM like interface to catalogues. Everyone: The RRS!   Ian: We should not solve this via pool, but via a proper tools  Markus: This is important for interoperation with the US. Ian: RRS defines the interface. Erwin should bring this up in the TCG.  Ian: RRS presentation is needed. Best in the TCG.\cf0  
\f1\b0 \
\ls7\ilvl0
\f0\b {\listtext	4.	}Good performance
\f1\b0    \cf4 -----------> general agreement but Dirk.\uc0\u8232 
\f2\i \cf0 Performace that privileges read access, up to read-only unauthenticated instance
\f1\i0  \uc0\u8232 
\f2\i if it helps.      \cf4  --------------> Different front end.  
\f1\i0  Flavia: Replication of read only. Dirk: Everything is there, GD should tell us when. \uc0\u8232 
\f2\i \cf0 The LFC should be highly optimized with respect to different kinds of queries,
\f1\i0  \uc0\u8232 
\f2\i bulk operations for file and replica registration should be supported.
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl220\sa20\ql\qnatural

\f0\b\fs22\fsmilli11400 \cf3 4.4 Grid Data Management Tools\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls8\ilvl0
\fs24 \cf0 {\listtext	1.	}lcg-utils available in production
\f1\b0 \
\ls8\ilvl0
\f0\b {\listtext	2.	}POSIX file access based on the LFN
\f1\b0  \uc0\u8232 
\f2\i The C/C++ API (gfal library) should be able to provide POSIX file access
\f1\i0  \uc0\u8232 
\f2\i based on the file LFN. This should include an efficient strategy for the
\f1\i0       \cf4 ---> LHCb: gfal takes the first replica Dirk: you can specify.  LHCB. you should implement a strategy to pick the best. Ian: What behaviour is asked for.  LHCb: The open should pick the most effective replica. Erwin: In edg you could pick different strategies, first, local, by best network connection. LHCb: Posix open. Lseek and best performance. The gfal has to pick best replica. Data will be local, but sometimes not, there sparse access to remote file. Best by xrootd.  Ian: You want to access the files local and remote transparantly via Poisix interface.  \uc0\u8232 
\f2\i \cf0 "best replica" choice in the context of a running job. The strategy should take
\f1\i0  \uc0\u8232 
\f2\i into account site location, prioritization of the different storage classes,
\f1\i0  \uc0\u8232 
\f2\i the current state of the networking, etc.
\f1\i0  \
\ls8\ilvl0
\f0\b {\listtext	3.	}File access API (gfal library) using multiple instances of LFC
\f1\b0  
\f2\i The basic file access API ( gfal library ) should be able to talk to several
\f1\i0  \uc0\u8232 
\f2\i instances of the LFC catalog to ensure redundancy for high availability as well
\f1\i0  \uc0\u8232 
\f2\i as load balancing for efficiency. \cf4  -------> Load balanced l
\f1\i0  \
\ls8\ilvl0
\f0\b \cf0 {\listtext	4.	}Reliable registration service
\f1\b0  \uc0\u8232 _ Supporting ACL propagation between storages and catalogs and bulk operations._ \
\ls8\ilvl0
\f0\b {\listtext	5.	}Reliable (bulk) file replica deletion service
\f1\b0  \cf4 --------------> ATLAS: different behavior, not reliable, and speed is not high enough. Flavia: This is about different interpretation of the standard. Ian: Internally there will be always different.  Markus: Please rephrase that this that you have the requirement to be able to delete X files in Y hours, or you might gain nothing. all: General agreement on this.\uc0\u8232 
\f3\i\b \cf0 Use Case:
\f1\i0\b0  
\f2\i delete all SC3 data (specify a set of files) sitting
\f1\i0  \uc0\u8232 
\f2\i on a storage element - a simple way to control that the deletion actual occurs,
\f1\i0  \uc0\u8232 
\f2\i with automatic handling of failures.
\f1\i0  \uc0\u8232 
\f2\i This should evolve into a standard cache turnover service(???)
\f1\i0    \cf4 -------> ATLAS: We discussed this, but is not yet a requirement\
\ls8\ilvl0
\f0\b \cf0 {\listtext	7.	}Staging service needed
\f1\b0    \cf4  ---------> LHCb: This should be done by FTS.  ATLAS: we want to do this in a T1 locally. (SRM-prepare to get)\uc0\u8232 
\f2\i a higher-level service to deal with staging of collection of files (datasets).
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 5. Workload Management\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls9\ilvl0
\fs24 \cf0 {\listtext	1.	}Stable and redundant service
\f1\b0  \uc0\u8232 
\f3\i\b ALICE:
\f1\i0\b0  
\f2\i Need a site specific configuration which contains a set of primary RB's
\f1\i0  \uc0\u8232 
\f2\i to be used by each VO (it can be one RB or more depending on the VO
\f1\i0  \uc0\u8232 
\f2\i requirements) and a second set of RB\'d5s which will be used in the case
\f1\i0  \uc0\u8232 
\f2\i the first set is down.    \cf4  --------->  ALICE: We have sites in Italy. Decentralizes the services: CNAF, Torino, ..... then all down we go to CERN-1, Erwin: The current system can handle a list. ALICE: But we need a list of lists. First line is a list with the most preffered first 
\f1\i0  \uc0\u8232 
\f3\i\b LHCB:
\f1\i0\b0  
\f2\i A list of RB's available for the VO should be defined and an easy or
\f1\i0       \uc0\u8232 
\f2\i transparent switching mechanism from one RB to another should be provided.
\f1\i0    -------> Ian: this corresponds to one entry point with many nodes behind it   Ian: We need to make this a scalable service. This has to be done in the new code base.\uc0\u8232 
\f2\i Ideally, a single RB end-point should be provided with an automatic load
\f1\i0  \uc0\u8232 
\f2\i balancing between the RB services behind. No loss of jobs or loss of the job
\f1\i0  \uc0\u8232 
\f2\i results due to temporary unavailability of a RB service should happen.
\f1\i0 \
\ls9\ilvl0
\f0\b \cf0 {\listtext	3.	}Capability of handling 10**5 short (>= 30') jobs in 1 day with <= 10 (approx) RB's.
\f1\b0  <br> 
\f3\i\b ATLAS/CMS:
\f1\i0\b0  
\f2\i Feature needed for SC4. The final short job number is evaluated
\f1\i0  \uc0\u8232 
\f2\i to be 10**6; thus the capability has to scale to 10**6 by summer 2007.    \cf4 ---------->    Ian: We need a service that can handle the load.  CMS: this is more for the developers.  
\f1\i0  \uc0\u8232 
\f3\i\b LHCb:
\f1\i0\b0  _~1Hz submission rate.\cf0 \
\ls9\ilvl0
\f0\b {\listtext	4.	}Efficient use of information system in the match making
\f1\b0  \uc0\u8232 
\f2\i Capability of sending the jobs to the sites where the input files are
\f1\i0  \uc0\u8232 
\f2\i present and having enough free CPU slots.
\f1\i0  \
\ls9\ilvl0
\f0\b {\listtext	5.	}Bulk job managing down to matchmaking and LRMS  \cf4 ----------> Erwin: What is meant by bulk  job matching?  Ian: This is fileblock matching Atlas: This is there and we test this. (DACS). You can specify that all the jobs of a set go to the same resource.
\f1\b0 \
\ls9\ilvl0
\f0\b {\listtext	6.	}Efficient input sandbox management (Caching of input sandboxes at sites ?)   Markus: Many jobs, almost identical sandbox. Erwin: The identical data, you can handle. CMS: 100k jobs with the same sandbox. 100 sites start. all hit the same node.  Markus: This is same problem with the condition db Erwin: This is a datamangent problem, we need a distribution. 
\f1\b0 \
\ls9\ilvl0
\f0\b \cf0 {\listtext	7.	}Latency for job execution and job status reporting should be proportional to the expected job duartion.  \cf4  ---------->Ian: lets skip
\f1\b0 \cf0 \
\ls9\ilvl0
\f0\b {\listtext	8.	}Support for different priorities based on VOMS groups/roles
\f1\b0   \cf4 ------>  Flavia: Global level. \uc0\u8232 
\f3\i\b \cf0 ATLAS:
\f1\i0\b0  
\f2\i This should be possible without relaying on a unique
\f1\i0  \uc0\u8232 
\f2\i centralized DB (gPbox)
\f1\i0  \
\ls9\ilvl0
\f0\b {\listtext	9.	}Fair share across users in the same group  \cf4  ---------> Discussion 
\f1\b0 \
\ls9\ilvl0
\f0\b \cf0 {\listtext	10.	}Interactive access to running job
\f1\b0      \cf4   ----------------> gridPeek is going a long way   CMS: Top. Ls, and peek at individual files needed.\uc0\u8232 
\f2\i For debugging and monitoring purposes
\f1\i0 \cf0 \
\ls9\ilvl0
\f0\b {\listtext	11.	}Computing Element service with direct access
\f1\b0     \uc0\u8232 
\f2\i Get the status of the computing resource and, in particular, the number
\f1\i0  \uc0\u8232 
\f2\i of waiting/running tasks for the given VO.
\f1\i0  \uc0\u8232 
\f2\i Submit, monitor and manipulate jobs through the CE service interface.
\f1\i0  \
\ls9\ilvl0
\f0\b {\listtext	12.	}Allow running special jobs (Agents) on a worker node to stear other jobs      \cf4  --------> LHCb: Pilot agents. This is an LHCb service that can handle other users jobs. This is the same as the trusted identity change service. 
\f1\b0  \uc0\u8232 
\f2\i Agents can steer execution of the jobs belonging to other users on the
\f1\i0  
\f2\i same worker node.
\f1\i0  \cf0 \
\ls9\ilvl0
\f0\b {\listtext	13.	}Allow for changing identity of a job running on the worker node
\f1\b0  \uc0\u8232 
\f3\i\b LHCb:
\f1\i0\b0  
\f2\i Interrogate the site policy service for permission to run a job of
\f1\i0  \uc0\u8232 
\f2\i a particular user.
\f1\i0  \uc0\u8232 
\f2\i In case of the positive answer, the new user proxy will be acquired
\f1\i0  \uc0\u8232 
\f2\i from the VO service for subsequent job operations.
\f1\i0  \uc0\u8232 
\f2\i The Agent job continues even after the user job execution finished.
\f1\i0  \uc0\u8232 
\f3\i\b ATLAS:
\f1\i0\b0  
\f2\i Using WMS to submit jobs doing data transfer on behalf of multiple users.
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 6. Monitoring Tools\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls10\ilvl0
\fs24 \cf0 {\listtext	1.	}Tools needed to monitor transfer traffic    \cf4  -----> Yes 
\f1\b0 \
\ls10\ilvl0
\f0\b \cf0 {\listtext	2.	}SE monitoring
\f1\b0  \uc0\u8232 
\f2\i Needed statistics for file opening and I/O by file/dataset from SE's.   \cf4 ----> abstract load figures\cf0  
\f1\i0 \
\ls10\ilvl0
\f0\b {\listtext	3.	}A scalable tool to collect VO specific information for global operations
\f1\b0    -\cf4 ----> CMS: Some way to get info about what the jobs are doing. We send a message from the job. Ian: This is fine, monaLisa or R-GMA\uc0\u8232 
\f2\i \cf0 Job status/failure/progress information
\f1\i0 \
\ls10\ilvl0
\f0\b {\listtext	4.	}Publish/Subscribe to logging and bookeeping and local batch system events for all jobs in the VO.   \cf4 ---->  R-GMA.  
\f1\b0 \cf0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 7. Accounting\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls11\ilvl0
\fs24 \cf0 {\listtext	1.	}Support for accounting, with site, user and group granularity (DGAS or equivalent)  CMS: VOMS groups. 
\f1\b0  \uc0\u8232 
\f2\i VOMS group information should be obtained from Proxy.
\f1\i0 \
\ls11\ilvl0
\f0\b {\listtext	2.	}Possibility to aggregate by VO (user) specified tag
\f1\b0       \cf4  --------->\cf0  \cf4 CMS: Like how much used by MC, how much used by Recon\cf0 \uc0\u8232 
\f2\i Application type, executable, dataset
\f1\i0 \
\ls11\ilvl0
\f0\b {\listtext	3.	}Storage Element accounting aggregated by datasets (e.g. PFN directory)
\f1\b0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 8. Applications\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls12\ilvl0
\fs24 \cf0 {\listtext	1.	}Address library conflicts with Middleware
\f1\b0  \uc0\u8232 
\f2\i Castor, LSF, POOL, DPM, etc
\f1\i0  \
\ls12\ilvl0
\f0\b {\listtext	2.	}Improvements/new features for the POOL File Catalog interface
\f1\b0  \uc0\u8232 
\f3\i\b ATLAS
\f1\i0\b0  
\f2\i Being discussed with POOL and LFC teams.
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 9. Deployment Issues\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls13\ilvl0
\fs24 \cf0 {\listtext	1.	}LFC global file catalogue available at CERN
\f1\b0  \uc0\u8232 
\f2\i Request coming from 
\f3\b CMS
\f2\b0  and 
\f3\b LHCB
\f2\b0 .
\f1\i0 \
\ls13\ilvl0
\f0\b {\listtext	2.	}Read-only mirrors of the central LFC service
\f1\b0  \uc0\u8232 
\f2\i Read-only mirrors should be available at a subset or all the T1 sites.
\f1\i0  \uc0\u8232 
\f2\i The mirror update frequency is of the order of 30-60 minutes.
\f1\i0  \
\ls13\ilvl0
\f0\b {\listtext	3.	}Each site should provide a Storage Element with an SRM interface
\f1\b0 \
\ls13\ilvl0
\f0\b {\listtext	4.	}Different classes of SEs
\f1\b0  
\f2\i Tier1 sites as well as analysis Tier2 sites should provide different
\f1\i0  \uc0\u8232 
\f2\i classes of storages with distinct SRM end-points:
\f1\i0  \uc0\u8232 
\f3\i\b MSS storage
\f1\i0\b0  
\f2\i (if available ) for non-frequently accessed data (archives);
\f1\i0  \uc0\u8232 
\f3\i\b Disk storage with write access
\f1\i0\b0  
\f2\i for production managers;
\f1\i0  \uc0\u8232 
\f3\i\b Disk storage with write access
\f1\i0\b0  
\f2\i for all the VO users.
\f1\i0  \uc0\u8232 
\f2\i A mechanism for choosing the SE at a given site with the above mentioned
\f1\i0  \uc0\u8232 
\f2\i characteristics should be provided.
\f1\i0  \
\ls13\ilvl0
\f0\b {\listtext	5.	}XROOTD deployed at all sites
\f1\b0 \
\ls13\ilvl0
\f0\b {\listtext	6.	}VOBOX deployment at sites
\f1\b0  \uc0\u8232 
\f3\i\b ALICE:
\f1\i0\b0  
\f2\i Needed at all sites
\f1\i0  \uc0\u8232 
\f3\i\b ATLAS:
\f1\i0\b0  
\f2\i Needed only at a central location
\f1\i0  \cf4 -----> ATLAS: Need it at all sites.  Markus: All sites including 5 nodes sites ATLAS: yes \uc0\u8232 
\f3\i\b \cf0 CMS:
\f1\i0\b0  
\f2\i Not needed
\f1\i0  \uc0\u8232 
\f3\i\b LHCb:
\f1\i0\b0  
\f2\i Needed at all T1 centers and selected T2
\f1\i0  \
\ls13\ilvl0
\f0\b {\listtext	7.	}VOBOX should be considered basic provided Grid services
\f1\b0  \uc0\u8232 
\f2\i VOBOX are provided as basic services with specific functionality.
\f1\i0  
\f2\i As such, it is the responsibility of site administrators to keep them up-to-date
\f1\i0  
\f2\i for what concerns the middleware services they provide.
\f1\i0  
\f2\i It is instead responsibility of ALICE to keep the experiment software installed
\f1\i0  
\f2\i on these machines up-to-date and to take care of possible problems that can
\f1\i0  
\f2\i occur when running the experiment specific agents.
\f1\i0 \
\ls13\ilvl0
\f0\b {\listtext	8.	}Each site should provide a Computing Element service accessible directly  \cf4  ----->LHCb: using the same interface.  But we need information access on the nodes.  
\f1\b0 \
\ls13\ilvl0
\f0\b \cf0 {\listtext	9.	}Support for short jobs
\f1\b0  \uc0\u8232 
\f2\i Every site should have dedicated queue for short (less then 30 min e.g. jobs)
\f1\i0  \uc0\u8232 
\f2\i so that those are executed with priotity. Job latencies should be proportional
\f1\i0  \uc0\u8232 
\f2\i to job duration.
\f1\i0  \
\ls13\ilvl0
\f0\b {\listtext	10.	}Standards for CPU time limits
\f1\b0 \
\ls13\ilvl0
\f0\b {\listtext	11.	}Support for queues with at least 2 different priority levels
\f1\b0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 10. Operations\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls14\ilvl0
\fs24 \cf0 {\listtext	1.	}Extend Site Functional Test to a heartbeat test for all major functionalities
\f1\b0  \uc0\u8232 
\f2\i Job execution,file transfers,storage access, etc.
\f1\i0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 11. VOBOX\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls15\ilvl0
\fs24 \cf0 {\listtext	1.	}CMS
\f1\b0  {\field{\*\fldinst{HYPERLINK "https://uimon.cern.ch/twiki/bin/view/CMS/CmsVoServices"}}{\fldrslt \cf6 \ul \ulc6 CMS VOBOX specs}}\cf6 \ul \ulc6         ----------> Flavia: I'll sync this with 9\cf0 \ulnone \
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls15\ilvl0
\f0\b \cf0 {\listtext	2.	}LHCB
\f1\b0  {\field{\*\fldinst{HYPERLINK "https://uimon.cern.ch/twiki/pub/LHCb/LHCbTaskForce/LHCb_VO-box.pdf"}}{\fldrslt \cf6 \ul LHCb VOBOX specs}}\
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 12. Castor standing open issues\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls16\ilvl0
\fs24 \cf0 {\listtext	1.	}Problem using Castor2 and SRM 'isCached'
\f1\b0  \uc0\u8232 
\f2\i Castor2 has different diskpools at the backend, but the SRM only sees
\f1\i0  \uc0\u8232 
\f2\i one of the diskpool. So a file is put onto a diskpool but is seen as
\f1\i0  \uc0\u8232 
\f2\i 'not being cached' by the SRM because it's checking the wrong diskpools.
\f1\i0  \uc0\u8232 
\f2\i Diskpools should either be transparent: provided that the copy between pools
\f1\i0  \uc0\u8232 
\f2\i is fast - or not transparent, but then visible/mapped somehow to the "grid" part.
\f1\i0  \
\ls16\ilvl0
\f0\b {\listtext	2.	}A User DN is mapped to one Castor pool only
\f1\b0 \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\pard\pardeftab720\sl260\sa20\ql\qnatural

\f0\b\fs26\fsmilli13200 \cf3 13. Miscellaneous\
\pard\tx220\tx720\pardeftab720\li720\fi-720\sl360\ql\qnatural
\ls17\ilvl0
\fs24 \cf0 {\listtext	1.	}xrootd interfaced with SRM
\f1\b0               \cf4  ------> Ian: this is well understood LHCb: we would like to use it Ian: should be taken up by TCG. Dirk: Why is there no one in the TCG from the application area ERWIN: We need to get them in for the focussed issues Ian: We need to start to draw up a list to start workshop like discussion.    \uc0\u8232 
\f2\i \cf0 xrootd is about to provide SRM interface. xrootd should be
\f1\i0  
\f2\i provided in production.
\f1\i0 \
\ls17\ilvl0
\f0\b {\listtext	2.	}CMS does not require Posix-like open of non-local SE's
\f1\b0 \
\ls17\ilvl0
\f0\b {\listtext	3.	}Hosting long-lived processes
\f1\b0  \uc0\u8232 
\f2\i Work on a standard set of secure containers? e.g Apache+mod_gridsite
\f1\i0  \uc0\u8232 
\f2\i as a site component? How to run agents using those services? As normal jobs
\f1\i0  \uc0\u8232 
\f2\i at the site?
\f1\i0  \uc0\u8232 
\f2\i Is it worth looking into the model of FTS with it's VO-specific agents
\f1\i0  \uc0\u8232 
\f2\i framework? Can the same principles be applied elsewhere? Is it possible to have
\f1\i0  \uc0\u8232 
\f2\i more documentation on this?
\f1\i0  \
\ls17\ilvl0
\f0\b {\listtext	4.	}Publishing experiment specific info
\f1\b0  
\f2\i Where should experiment specific info be published? BDII, R-GMA, ...?
\f1\i0  \
\pard\pardeftab720\sl360\ql\qnatural
\cf0 \
\
\pard\pardeftab720\sl360\sa120\ql\qnatural
\cf5 \cb5 \
\pard\pardeftab720\sl360\ql\qnatural

\f0\b \cf0 \cb1 Major updates
\f1\b0 :\
-- Main.flavia - 29 Nov 2005\
\
\pard\pardeftab720\sl360\ql\qnatural

\fs22\fsmilli11400 \cf7 Revision: r1.4 - 01 Dec 2005 - 10:00 - Main.flavia\
\pard\pardeftab720\sl360\ql\qnatural
{\field{\*\fldinst{HYPERLINK "https://uimon.cern.ch/twiki/bin/view/Main/WebHome"}}{\fldrslt 
\fs21\fsmilli10800 \cf6 Main}}
\fs21\fsmilli10800 \cf0  > {\field{\*\fldinst{HYPERLINK "https://uimon.cern.ch/twiki/bin/view/LCG/ManagementBoard"}}{\fldrslt \cf6 ManagementBoard}} > {\field{\*\fldinst{HYPERLINK "https://uimon.cern.ch/twiki/bin/view/LCG/TaskForces"}}{\fldrslt \cf6 TaskForces}} > SummaryOpenIssues
\fs24 \
\pard\pardeftab720\sl300\ql\qnatural

\fs21\fsmilli10800 \cf7 Copyright \'a9 1999-2005 by the contributing authors. All material on this collaboration platform is the property of the contributing authors. \
Ideas, requests, problems regarding TWiki?\
\
\
\
\
\pard\pardeftab720\sl360\ql\qnatural

\fs24 \cf0 Ian: we need another meeting to finish it up. Not next Friday. \
\
Better before 16th of December. \
\
Ian: I send a list with topics that have to be written up. Please sent contributions. \
\
We have to meet asynchronously. \
\
Ian: We need to prioritize. I'll start and propose something as soon as I have something.\
}