|
The graph above shows the total Million SPECint 2000 (MSI2K)* accessible to the OSG infrastructure as proposed in the initial project execution plan and as measured in March 2007 and February 2008.
The 2007 measurement was made through a manual survey of processing resources. Only 40% of the sites responded and a projection to 100% was then made. (This report was sent to the JOT and Council last year and is available on request from osg-webmaster at opensciencegrid.org. The 2008 measurement was made based on the number reported by the OSG Information Service (Generic Information Provider). 63 resources (~85%) reported and 3 by-hand corrections had to be made. To-date, we haven't independently validated the result. We know that the number fluctuates for specific activities, for example, for the ATLAS analysis challenges and through guaranteed commitments during the 2007 D0 reprocessing effort.
Please note that this is the MSI2K reported as accessible to the shared infrastructure on a particular day. Local policies govern the use of the resources, and only a percentage of the resources are usable by organizations other than those that own them.
The maximum current usage of OSG is 16,000 CPU Wall Clock Days/Day, with 80% of the sites reporting to the accounting system. The usage has dropped somewhat in the last week due to a hiatus in ATLAS event production and analysis.
~ OSG EB
* From the OSG Project Execution Plan: "Conversion from MSI2K is 1500-1700 SI2K/CPU in 2008."
|
  |
|
 |
Ruth Pordes |
We held our annual All Hands Consortium Meeting, followed by the semi-annual Council meeting, earlier this month. At the end of his two year term, Bill Kramer stepped down as Chair of the Council. We have greatly benefited from his wise and experienced leadership during the past two years and look forward to his continuing involvement in the work and direction of the consortium.
We are further fortunate that Paul Avery and Kent Blackburn stepped forward and were elected as the new Council co-Chairs.
At the All Hands itself, there was lively discussion throughout and the exchange of a lot of information and ideas for the future. I really wish we had recorded at least the plenary sessions so we could share this with the rest of you!
There was a whirlwind of activity throughout the meeting: We held a Biology Applications Workshop, a Campus Infrastructure workshop, the next in our series of Site Administrators meetings and ATLAS and CMS co-located meetings. There were discussions from our science users, briefings on OSG's status and reports of worldwide cyberinfrastructure, outreach and education. Even at the end of three long days people listened to three alumni from our grid schools who gave us their feedback and explained their plans for using distributed computing in the future.
The first testing phase towards the release of the OSG 1.0 software has now started on the Validation Test Bed, with testing on the Integration Test Bed (ITB) to follow shortly.
In this release of OSG, the VDT package will no longer include a version of OpenSSL. Instead, it will rely on the pre-installed version of OpenSSL on each site. We have done a lot of additional testing of the VDT with this change and will be working closely with the sites to ensure there are no unanticipated problems.
~ Ruth
|
|
|
