CERN and sustainability

Back to index page

Next article: We are farmers, not hunter gatherers

Literally ‘the coolest place on the planet’, things are really heating up at CERN as final preparations for data collection take place at the Large Hadron Collider (LHC). The Worldwide LHC Computing Grid (WLCG) project has recently completed tests that, in the author’s opinion, have been the most demanding and convincing production demonstration of grid computing to date.

To process the massive quantities of data that will be generated when the LHC is operating – some 15PB annually – a worldwide computational and data grid has been built. The LHC is expected to operate until around 2020, by which time a significant upgrade to it, and indeed the entire accelerator infrastructure of CERN, will have taken place. Long-term sustainability is crucial to CERN’s operations.

Sustainability is no easy challenge. It will almost certainly require adaption to major shifts in technology – if the past is any judge of the future. Together with a large number of ‘Tier1’ and ‘Tier2’ sites worldwide, including many in the UK, the WLCG has completed an extensive series of tests to demonstrate readiness for data taking and processing. This has required worldwide collaboration between several hundred institutes and (at least) three distinct grid environments (primarily EGEE and OSG). The tests have shown sustained import and export of data between the required combinations of sites and at the necessary rates (several Gbs, for days at a time). They also demonstrated the processing of massive numbers of production and user jobs, and a high degree of reliability and usability.

But change – as always – is on the horizon. The current funding model for Grids is evolving towards a long-term, sustainable e-infrastructure, like the European Grid Initiative. At the same time, potentially new paradigms are emerging. ‘Cloud Computing’ could open the door to truly ubiquitous computing, as long as it can demonstrate successful examples of its use. This could lead to a situation similar to that when the Web burst out of the research arena and developed into a tool used by virtually everyone. However, the benefits can be expected to be much greater – given that there is essentially unlimited freedom in the type of algorithms and volumes of data that can be processed by a Cloud.

One of the positive aspects of today’s Grid deployments is that they are breathing life back into the internationally distributed research institutes that form the Grids. Enabling scientists at these institutes to have equal access to data and play a full role in the collaboration avoids ‘brain drain’, because researchers no longer need to be based near a central research institute. Local industry is stimulated by the local spending of funding, which is something that funding bodies find very attractive. Of course, these benefits occur in addition to the usual benefits derived from technology transfer and spin-outs that result from the presence of vibrant scientific institutes in the region. Ideally, funding bodies should obtain similar or equivalent benefits from an alternate computing paradigm – or even greater benefits obtained from more efficient use of the funds provided.

Adapting to change is a fundamental component of any long-term IT strategy. Doing this in a non-disruptive fashion, whilst offering a 24x7 peta-scale worldwide computing environment is a challenge – or rather, an opportunity.

Jamie Shiers, LHC Grid, CERN.

www.tinyurl.com/6amhhe