End-User Analysis Scenarios

A working group has been created in the MB to establish what kind of analysis scenarios at CEN would be feasible.

More details can be found in the minutes of the corresponding LCG MB meeting (08. July 2008).

MB Minutes : https://twiki.cern.ch/twiki/pub/LCG/MbMeetingsMinutes/LCG_Management_Board_2008_07_08.html

MB talk : end-user_analysis_disk_space_v2_7July2008.ppt

The idea would be to establish by the beginning of September what kind of options we have for an implementation of this new analysis infrastructure in the time frame of October and also on a longer time frame like spring 2009.

Working group members

  • ALICE : Andreas Peters and Fabrizio Furano
  • ATLAS : Kors Bos
  • CMS : Stephen Gowdy and Ian Fisk
  • LHCb : Patrick Koppenburg
  • IT/FIO group : Miguel Santos
  • IT/DM group : Dirk Duellmann and Alberto Pace
  • Chair : Bernd Panzer-Steindel

ATLAS and CMS have made an official request for an analysis infrastructure and already provided some requirements for this :

ATLAS request note : * ATLAS_Memo-on-Userspace.pdf

CMS draft requirement note : * cmscafRequirements.pdf

A first draft of the possible scenarios taking into account the known requirements and boundary conditions has been produced :

As a quick summary one can make two very extreme scenarios for 500 users and calculate a rough cost envelope for a period of 3 years.

  • ideal from the user point of view (centrally managed)
      world-wide accessible distributed file system with automatic quotas, load balancing, security and backup
      500 TB backed-up AFS space plus 100 TB AFS space for data replicas
      --> 3 MCHF for AFS + 1 MCHF for TSM backup + 2 FTE extra for operation  
            = 4 MCHF + 2 FTE

  • ideal from the operations point of view (completely user managed)
      local disk space on a notebook plus cloud computing space 
      --> 500 CHF for 1 TB mirrored USB disk + 2500 CHF for 200 GB Amazon S3 space and 10 GB traffic per day
            = 1.5 MCHF        no extra FTE

Summary of the boundary conditions

  1. available hardware at the current time
  2. timeline to buy additional hardware
  3. additional costs
  4. availabe software with the needed functionality , time frame for needed adaptations (minimal developments)
  5. but most important is the operational aspect, which would mean 'reallocation' of FTE's from other tasks to this new facility

In the mean time meetings between the Castor development team and ATLAS and CMS took place to establish the Castor plans and possibilities in the analysis area :

Feedback to the first draft has already been received from different persons :

* Collected_feed-back_analysis_working_group.doc:

A face-to-face meeting between the working group members and invited guests took place on the 27th of August.

Based on this meeting and all the other feedback a proposal for the short term solution was send to the IT physics management, as the required service is heavily dependent on the existing operational effort in IT.

* Quick_summary_and_conclusions_01Sept2008_v2.doc

The document was discussed with the responsibles in the data management (DM) group, the operations group (FIO) and the physics group meeting. The conclusion from IT is to offer an analysis storage facility based on CASTOR. The following document shows the details as part of an SLA agreement.

* Castor_proposal_by_Olof_Barring.doc

The proposal is to offer the service by the first of October and improve the needed analysis functionality continously during the last quarter of 2008.

The medium term plan is provided in a document from the DM group.

* 2008-09-08-Plan-for-LHC-data-analysis-v2.doc

The focus of the working group was on the short term deployment and medium term plan for an analysis storage facility at CERN. As this has now been done with the mentioned plans, the working group will stop now.

-- BerndPanzersteindel - 16. September 2008

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2017-01-05 - BerndPanzerSteindel
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback