Friday, November 5, 2010

Wrapping up at the e-Infrastructures meeting

It’s been a busy couple of days here at CERN in the Globe, and as we in e-ScienceTalk have been involved in the organisation of the event, we’ve been delighted to see so many projects represented. It’s difficult to pick out highlights from such a rich set of programmes – as Enric Mitjana, the call coordinator pointed out this morning, of the 107 proposals received in Call 7 last year, 38 have been funded, so the quality is very high. There are 80 projects active at the moment, but we heard from some of the newly funded today.

I introduced e-ScienceTalk this morning, as well as giving a very brief overview of the many excellent coordination and support projects funded in the Call – you can find out more at We are starting to work with many of them already – someone commented that they were waiting to hear which projects we DIDN’T want to work with, but in e-ScienceTalk, we really are keen to talk to everyone! Later we saw a series of snapshots from the Virtual Research Communities projects, and those funded in the Software Simulation and Services topic.

Bernard Marechal introduced GISELA , describing the project as the new baby - grandaughter of EELA and daughter of EELA-2, with a Latin American and Caribbean parentage. GISELA will be working with NRENs and NGIs to encourage CLARA and the Latin American NRENs to support research infrastructures in the long-term. They are focussing on working at 3 levels – local, national, regional.

Steven Newhouse of also gave a thought-provoking overview of the six major distributed computing infrastructures that have kicked off in the last few months: EGI-InSPIRE, European Middleware Initiative (EMI), Venus-C, StratusLab, European Desktop Grid Initiative and the Initiative for Globus in Europe (IGE). Working together with coordinating standards project SIENA, the six projects have been doing some assessment of the real barriers to broader adoption of e-Infrastructures. Up til now for example, the sustainability of e-infrastructures has not been clear, a problem if you are a researcher worried about your data long term. This is now partially addressed by the new EGI structure, based on NGI and EIRO stakeholders (including CERN and EMBL). To date, advances in ICT have matched data analysis needs, but this is changing with the expected data deluge from the ESFRI projects. The fact remains that computing beyond your desktop is difficult for many – but virtualisation and new computing models should help simplify this. Service offerings do not always match the needs of new users – we have to meet their needs through a flexible federated production infrastructure. With the advent of clouds, it’s also important to compare up front the value, benefits and costs of public and commercial offerings.

So this is all food for thought for the projects over the coming years – but while cogitating on these barriers, they also realised some of the unique selling points that shouldn’t be forgotten. Often the compute resources are near the storage resource, co-locating processing near the data. Moving large quantities of data around commercially can be time-consuming and expensive. For e-Infrastructures, the high speed networks are ‘free’ at the point of use. The projects have more than ten years of trust built up in the community – sharing resources in a trusted environment is now second nature. In that time, the community has also built up close links to data-oriented users and have developed a close understanding of the patterns in the ways that researchers use data.

So with a couple of sessions still to go, we will soon be leaving the warm wooden bubble of CERN’s Globe and re-emerge back into the fresh air of the Genevan autumn and the outside world. For us as a project it’s been a great opportunity to speak to the members of our new e-Infrastructure family – we look forward to the next family reunion!

No comments: