Notes for the ITC-Research Computing Standing Committee Meeting

October 13, 2008 at 10:30 AM
2015 Ivy Road, 1st Floor Conference Room

Members: Alice, Andrew, Hamp, Jim, Joe S., Mark, Mike, Robin, Terry, Tim S., Tom and members of Systems, Research Computing and ACHS as desired

Attending: Alice, Andrew, Hamp, Jim, Joe S., Mark, Mike, Robin, Tim S., Tom, Katherine, Ed
Chair: Tom
Recorder: Alice

I. Corrections to minutes from last meeting on September 8, 2008? (Click to review the agenda from this meeting)

II. Old Business:

• Update on CSAC, NSF PetaScale Track 2 meetings, and related.
• Andrew gave an overview of his NSF Petascale Track 2 proposal for a cross-campus grid – (Note: the CSAC group devoted a lot of their most recent meeting to this draft proposal): This is an infrastructure proposal targeted at high-throughput computing – and is a response to NSF 08-573 for “a pool of loosely coupled grid-computing resources”. The proposal is due on 11/28/08. The proposal for this data and compute grid highlights our differentiators, including: standards-based, simplicity and ease-of-use – with proposed response themes of: standards-based; simplicity; security; availability; applications; accounting; and the option to get resources by ‘rolling over to the cloud’ (e.g. Amazon cloud/EC2). The proposal strawman is to have UVa lead a team of 4-6 universities. Since being a partner involves an institutional commitment, James and Mike will help with this – some possibilities include Virginia Tech, Clemson, JMU, UCSD, maybe Tennessee, and Texas Tech. UVa could benefit with greater visibility, access to the grid, and it could help to create a test-bed environment.
• This was followed by further discussion about what a cross-campus grid might mean to our group – and Andrew gave a fuller description of his draft proposal. One risk going forward is that we/UVa may not have as many computers in public labs as we do now – especially after 3 years. This might partly be mitigated by new machines being more powerful – and also raises the importance of a strategy to recruit/incentivize more participation from faculty/staff/student computers. We could start with how to recruit staff machines and harvest cycles using LSPs in departments. This would be more complicated than using public labs where there is a single build – so one would have to opt in and install something -- and we would need an accounting structure – but it would be good to develop a strategy for this. There would be practically no marginal costs for those who donate their computer cycles. We are getting data about existing lab usage.
• Other risks include: software fails to scale; security risks (need to audit like crazy); institutions start charging for electricity; the grid turns out to be too hard for users to use.
• What about file storage? Can make it look like a mounted file system.

• Storage needs assessment -- LSP dialog, survey (Joe)
• Although there was limited departmental participation, the Departmental Storage Report does give voice to department concerns and encouragement to ITC to pursue other storage options. Joe sent the SharePoint url for the report to the group.
• Given the current budget situation it remains to be seen if we can do anything about providing more/different storage.
• Cluster node purchase by faculty – revisit
• With ETF funds currently frozen, SEAS is trying to find alternatives for Michael Shirts (in ChemE – who needs fine-grained parallelism).
• A couple of other faculty (Martin Wu/Biology, Center for Public Health Genomics) are interested in memory.
• Possibility of acquired "gently used" systems
• Andrew will call JP Morgan
• Communication with potential benefactor: None.
• PBS Pro license purchased?
• Yes – purchased – but we need Flexlm licenses. We will go to 9.2 across the board by January.
• Update on new machine-room building
• We have the estimate from the consultant and Facilities Management – and it is $1 million over what the BOV approved – now we are figuring out how to deal with this – may need to go back and get permission to spend more. The building can be sized so that it can be doubled.
• Transition Plan for research support (with library) -- communication with users?
• Will wait for an update on this until Terry can attend.

III. New Business:

• Any other Items?

IV. Adjourn by 11:45 and next meeting

• Next Scheduled Meeting of ITC ResComp Standing Committee: November 10, 2008 in ITC-2015 Ivy Road, Room #102 (First Floor Conference Room)
  Please send suggestions, additions, corrections to: Tom or Alice