HiPCAT Meeting minutes for 3-1-2005

HiPCAT Meeting Agenda March 1, 2005

Tigre update—Phil Smith Parallel computing classes—Pat & Phil Grid activities—All (SURA, OSG, local grids, etc.) Institution updates—All

Meeting minutes:

Institutions attending: BCM, TTU, UT-Arlington, UT-Austin, TAMU, UH (Rice tried to attend but could not get a phone line). We had the most successful teleconference yet. All sites were that tried were able to connect by InSORs. It seems that all the noise canceling equipment is working well. The major problem encountered this time was the lack of dial in lines. It seems that InSORs only supports a few lines (perhaps 6). I will have a back-up dial in at TTU if we cannot get more lines from InSORs.

The Tigre project is progressing. It seems that we should have all institutions sign off by the end of March. UTEP, UT, and TTU are cooperating on producing some educational materials for parallel computing. We should have something to share with the group in April.

There are many grid activities going on. UT is involved in Teragrid, SURA, GridChem, United Devices, Condor, and TIGRE. TTU is involved in Tigre, Condor, Avaki, OSG, and SURA. UTA, UH, Rice, TTU, and TAMU are involved in THEGrid as well as several grid projects connected to various physics projects. Rice was the lead on GrADS and is the lead on VGrADS. Both major NSF funded multi institutional grid research projects. UH was/is a partner on both projects.

TAMU Kingsville joined the meeting and reported on their activities in HPC and grids. In particular, they have a significant interest in Environmental Modeling and Computing. UTA won a significant grant to be a tier 2 center for the LHC grid project. Alan Sill spoke about Open Science Grid and Virtual Organization Management Software (VOMS). He mentioned that VOMS is replacing grid map files as the way to keep track of virtual organization membership on the grid. EDG (LCG) and OSG have both adopted it. Alan thinks that the Asia-Pacific Grid (APGrid is also), but VO membership and certificate approval and sharing mechanisms are a different thing. Alan thinks the best way for this to proceed is for people to join the relevant working groups (e.g. Policy, Security, etc.) of OSG and to make their needs and preferences known. As a large regional grid effort, they would certainly get a lot of weight and attention in the OSG scheme. But in that OSG scheme, they would be a regional VO that has its own membership criteria - and hence the need to control and catalog that membership, for which VOMS and related tools were designed.

Alan added that people could continue their development, but be participants in the developing OSG process -- the best of both worlds. Alan noted: importantly, OSG is NOT a simple single grid! It is a framework for grids, with different organizations using the same or similar software, but controlling their membership and sharing their resources separately from other organizations on the grid.

BCM (David Steffen) discussed plans for the inperson meeting in May. He requested that schedules be sent to choose a date. Subsequently he has proposed May 27 as the meeting date. There will be a "Baylor Bioinformatics Day" on May 26 to which all HiPCATS are invited. More information will follow.