• Projects & Working Groups
• Scientific Computing
• E-Infrastructure for E-Science
  • News
  • Details
  • Projects
  • Submission
  • Download
  • FAQs
  • PWEG
  • Contact
• Cloud Computing
• Learning Infrastructure
• E-Identity
• ICT Law WG
• Mail WG
• Media WG
• Network WG
• Procurement WG
• Security WG
• Storage WG

Grid session

moderator: Andres Aeschlimann; protocol: Placi Flury

Projects presented:

1. reports and plans of Swiss Grid projects (summary; see also project ideas below)
2. SKIP (Sigve Haug; SKIP Blueprint)
    

Discussion:

Q. Martin, web-interface for global access? Demand from user base. -> what's the scope? Thomas has same issue with R.
Sergio: technically support isn't a problem. Ideas to group communities, in order to provide better support (and also estimate efforts). 
Thomas: R-portal may be better to attract users (advertise 'we have an R portal') then to go to individual users.  -> proposals should focus on attracting new/ establishing existing user bases.
Martin: that's often done in biology. Done very openly. examples of funding: France, EU, US NIA.., open to anyone. -> idea not to replace these things. Just provide an interface to e.g. SMSCG and to provide support.

ECGT (cancer related project), which is also using R, portal-like.

Montpellier model: as more users, as more EU money/funding.

Some bio-computing groups focus on cloud now and are no longer interested in Grid.
 

Project ideas:

- Selectome: short overview of application and project status.
Contacts with SMSCG established. Optimized version to be ready soon (eg. to be installed on WSL cluster).
Q. size of CPU time. -> average 20 minutes. 
Q. installation efforts etc. -> technical disc. on compilation settings etc.
Ideas for extension are around. Submission planed.

- SMSCG project targets at 2012 extension. Thomas asked for scope of extension. Sergio tells the last two scopes: 1 phase much time invested (more then planed) in applications. Difficulty to harmonize it with tasks
of operating infrastructure. SMSCG-II restrained to pure infrastructure, application as separated but affiliated projects.
SMSCG-II extension: currently CHIP, SKIP, EGI etc. are involving/consisting out of international efforts. SMSCG becomes national infrastructure. Idea is to align with European efforts.
Accommodation of use cases and supported applications might blow up scope.

Sigve asks how many arc sites on GOCDB? Few are.  What's the gain? May become too expensive for some sites.

- Multi-Site ARC Project (Marko Niinimäki; started last October, duration 12 months)
Extending SMSCG by integration of cluster and applications. 2 clusters. 3 applications. Large collaboration already inhouse.
Objective to apply for continuation.

- Sigve: we are thinking to move towards virtualization (possibly a virtualization project).
Sergio tells about VM-Mad project. Align with it to avoid duplication of work

- Sigve: Idea of cloud back-end for ARC. Sergio refers to some problems (eg. control of Amazon cloud).

- Data management project:
Sigve updates on status of project. Some storage space offered to SMSCG. Project ends in June. Sergio asks whether it can become a core service for SMSCG? Not yet thought about that. Evaluation of image-based classification project will help taking this decisions. -> continuation/extension depends on that.