Notes from SURAgrid In-person Meeting
1201 New York Avenue, Suite 430 * Washington, DC

February 22-23, 2006

The focus of this meeting was on the deployment of specific applications and further definition of a predictable and commonly useful application environment.

Attending:

SURA - Mary Fran Yafchak, Gary Crane
GSU - Art Vandenberg, Victor Bolet
GMU - Phil Yang
UNC-C - Barry Wilkinson
UKY - Vikram Gazula
ODU - Mike Sachon
TTU - Jerry Perez
TACC - Ashok Adiga
[Independent] - Judith Utley
MCNC - Steve Thorpe (via video conference for ADCIRC presentation)
UMich - Shawn McKee (attempted portions via telecon)
ULL - Denvil Smith
Vanderbilt - Paul Sheldon
LSU - Hartmut Kaiser, Archit Kulshrestha
Tulane - Rene Salmon
TAMU - Steve Johnson
UAB - John-Paul Robinson (attempted portions via telecon)
UAH - Sandi Redman

Not represented: UVA, UArk, UFL, GPN, OleMiss, NCState, SC

Wednesday, February 22, 10 a.m. - 5 p.m.

1) SURAgrid Application Environment - Ashok Adiga See attached: SURAgrid-resources.ppt

Discussion:
- Need to address account management & defining/naming SURAgrid users in order to make further progress beyond manual setup of individual user accounts on pre-identified resources.
- Should address Globus compatibility sooner than later and plan for when to move to Web services since some information in the Resource doc (and related procedures, scripts, etc.) varies for pre-Web and Web services implementations. Also, need to provide clarification somewhere for apps/users as to what level of compatibility they can expect across resources that may be using different versions of Globus. For this, could be addressed in the document that is being planned as a companion to the Adding Resources doc -> Running Your Application on SURAgrid.
- For verification of resources - Initial (current) approach is to simply ask sites to verify base level functionality, monitor for ongoing operation themselves, and supply publishable contact info for support staff. Ashok is working on a script to automate some of this.
When that happens, can give resource owners the option to open a test account for the script to be run regularly (with notification to support when problems are found), or to modify the script to run themselves against their own resources. Some of the verification could also be handled by the portal, which is already regularly/automatically pinging to see if resources are up or down.
Also have a growing list of other detail we would like to see added to the portal. Will begin a "wish list" and check with TACC following the meeting.
- Would like to identify some mechanism(s) for posting notices related to operation and support of SURAgrid resources, also provide means of interaction for problem-solving. Could begin with listserv, perhaps move to Wiki or other similar forum. Could also consider developing a Troubleshooting Guide.
- Have a growing list of detail we would like to see added to the portal. Will address this in a future (soon) SURAgrid call.

2) Review of Applications Underway - (What the application is, why it's important to grid-enable, description of the grid-enabling that's been done so far, discussion/detailing of next steps towards SURAgrid deployment, steps to be undertaken on Day 2 or plans for beyond)

Discussion points:
a) UNC ADCIRC - Steve Thorpe, MCNC (via VC) - See attached: UNC_ADCIRC_SURAGRID.ppt

b) Multiple Genome Alignment - Art Vandenberg, Victor Bolet - [presentation not received; GSU - please send to list when available)

c) ENDYNE - Jerry Perez - See attached: Endyne Sura Jerry Perez 2006.ppt - Endyne-Sura-JerryPerez-2006.ppt

d) SURAgrid CS Lab - MFY, Barry Wilkinson - See attached: SURAgrid-teaching-lab.ppt, Wilkinson-SURAGridpresentation.ppt

e) UCoMS - Archit Kulshreshtra - See attached: UCOMS_SURAGrid.ppt

f) SURAgrid Application Environment - Ashok Adiga

These applications (minus the CS Lab, which is still in concept stages) are now running or nearing verification on a subset of SURAgrid resources, with specific steps identified to assist in running on more.

*SCOOP/ADCIRC - Initial SURAgrid sites: TACC, UKY, TAMU, USC, ULL, LSU, Tulane *ENDYNE - Initial SURAgrid sites: TTU, GSU, TACC, TAMU *Multiple Genome Alignment - Initial SURAgrid sites: GSU, USC, UVA, TACC *UCoMs - Initial SURAgrid sites: LSU, TACC, GSU

Thursday, February 23, 8 a.m. - 3 p.m.

1) Regional Grid-related Activity - Gary Crane highlighted SURA's interest in solidifying and expanding SURAgrid and reviewed a corporate partnership proposal with IBM that is currently under exploration (see attached doc file for additional background). Jerry Draayer, SURA President, also joined at this time and added his endorsement of the SURAgrid effort and concept, thanking current participants for valuable work in making the concept tangible. Jerry also emphasized SURA's mission of enabling science and engineering, noting that SURAgrid is an initiative with clear potential to advance this mission and change paradigms in collaborative advancement of scientific knowledge.

2) "Deployment Fest" - Lab-style session all day.

Progress highlights:
- People worked individually and in ad hoc groups throughout the day on both running the applications that had been presented and also on bringing resources "further up" - in line with the more complete view of readiness that is being described in the Adding Resources to SURAgrid doc.
- Since folks left at various times and many were working on things right up until they had to go (and some intending to jump right in again once they got back!), there wasn't clear status on resource or application advancement at the end of the meeting. Ashok is currently gathering status from sites that were actively working on things, eventually progressing to all sites for a fresh view of the overall SURAgrid. When you receive a message from Ashok about this, please answer as quickly and fully as you can - thanks! (BTW - This information will also be used to refresh the sorely outdated Table of SURAgrid Elements that is on the main page of the SURAgrid Web.