(Draft) Agenda

General Schedule

Thursday

08:00

continental breakfast (provided)

08:30

start

09:45-10:30

break

12:15-13:15

lunch (provided)

14:30-15:00

break

17:00

finish, dinner on your own

Friday

08:00

breakfast

08:30

start

09:45-10:30

break

12:15-13:15

lunch (provided)

15:30

finish

Homework

• where is the problem
• what would components of an architecture be to solve end to end performance problems?
• prepare quick summary of own work (15 min max) for presentation

Annotated agenda without specific times

• First, discussion of scope: make sure we agree. In particular, is there a line between this (network performance) work and "grid" or "middleware" diagnostic measurement architecture.
• Discussion of outcomes. We intend to present some expected outcomes here, with another discussion at the end of the day.
  • A whitepaper
  • A (set of) architecture(s) that can be refined later
  • How existing projects fit into architecture
• Top-down: get a feel for architectures.
  Short presentations on existing proposed architectures. (15 min each)
  • Internet2 End-to-End pipes
  • GEANT phase II work
  • Middleware diagnostics (in particular, the Internet2 Middleware End-to-End Diagnostics work just underway)
  • Security considerations
• Bottom-up: commonality in existing projects
  Short presentations from implementers on existing projects (10 min each)
  • AMP & PMA
  • SLAC monitoring work
  • Pipes (including database/outreach work)
  • NIMI
  • Advisor
  • PSC work TAAD, new projects
  • ANML work (Steve Wallace's IU lab)
  • LBL work: SCNM, NTAF
  • ...
• Identify common themes in work.
• Discussion of "use cases" -- different people's view of the problem. See different people's interpretations, come up with a single consistent (set of) view(s).
• Discussion of different views of architecture. Common themes? Want, in particular:
  • common cases
  • unique cases
  • holes
• Look for single architecture. However, aim is not to find single thing to force down people's throats, but see if we can agree on common blocks.
• Discussion of the open research areas (starting with the holes?)
  Identified already
  • security/authentication: what is the right model?
  • resource discovery [how find boxes]
• Relationship to other groups
  • Standards
  • Areas: Grid, Middleware, Security
• Discussion on near-term things that we can do for each other: building our own community
• Recommendation on infrastructure support [pure research projects tend to die off]
• Discuss how to produce outcomes.