From Dr Bruce Becker, SAGrid coordinator:
Dear JRU and SAGrid Ops members
Please see the mail below from Roberto regarding the recent CHAIN Interoperability test in which SAGrid participated. There is a short youtube video as well: http://youtu.be/0R_1Wvgcgm0
SAGrid sites ran a few hundred jobs during the demo, at UJ, UFS, UCT and CHPC. Unfortunately, SAAO and NWU just missed out due to technical problems at the sites; Wits is entirely occupied with ATLAS jobs, which in itself is a good thing... So, this is a very good reflection on the state of our infrastructure.
But think about what this means for users ... This is a major step in the direction of reducing barriers to entry for users - users could log into this gateway using existing credentials and never have to use a personal certificate. Even if our infrastructure was completely nuked, they could still run their jobs anywhere in the world - they wouldn't even have to know why jobs went to a certain place because this is configured at the level of the application portlet itself. The test includcredentialsed infrastructures consisting of:
--> HPC clusters
--> Blue-Gene supercomputers
--> peer-to-peer pc clouds (similar to SETI@home)
and ran on 6 different kinds of middleware in 5 continents! The downside is the limited set of applications that have been ported to the gateway. These are anyway available on the grid, but you need the CLI to run jobs.
The solution to this, in my opinion is to promote the software stack that we're using - all standards-based java - and build a hacker community who can churn out portlets for new applications and extend the functionality of the virtual research environment within LifeRay. I'm trying to organise a hackfest at AIMS in January together with the SAAO to get things going... any other ideas are welcome!
The gateway will remain functional for a while still, so I suggest you go ahead and try it if you haven't yet:
http://science-gateway.chain-project.eu - you may even unwittingly do some real science
Cheers !
Bruce