Meeting Notes for 16 September 2009

General

• SWIFT has money for maintenance work
• Get more grid power - not just a grid of one (cluster)
• Tuesday, Mihael+Phong meet on 22 September at ND (arrive 11-12 Eastern Time)

Cosmic

• Mihael is playing with a computer tool for the cosmic ray detector: (Image)

LIGO

• Bluestone has SWIFTification code but is presently used.

CMS

• OGRE - what's the plan for grid infrastructure
  • We build a specific file structure, pick off the existing dimuon triggers
  • We are likely going to use a Tier 2 center for the CMS data
  • Right now the system runs only on one CPU
• Main ROOT cost is parsing the data (not reading it in!). Several seconds per file.
  • Test-beam data is ~80 files but not usually used in max.
  • Splitting files? Might lead to serious duplication.
• Plot <-> e-Lab bridge needs to be implemented in OGRE
• Non-viewable (in browser) image types
• Data needs to be controlled by someone at CMS
  • Implicit in the skimming?
  • In case there is an issue - we need to be able to pull and regenerate data.
    • What do we do with existing plots and posters?
    • Timestamp them?
    • Notification when data has changed?
    • Changelogs - insert into plot metadata?
  • Question - are we going to let the users discover nifty things - but also restrict them from the good data?