Whale watching at CHEP, eh?
Wed 26 Sep 2007
More than 20 GridPP members were joined by nearly 500 other high energy physicists, tens of totem poles and several orcas earlier this month at CHEP in Victoria, Canada. The conference brought together those working on computing for the LHC and other ongoing particle physics experiments, and included a look at plans for the International Linear Collider.
|
 |  |
The weekend before the conference saw a Worldwide LHC Computing Grid workshop, including panel discussion on site and experiment readiness and sessions for each of the experiments to consider their dress rehearsals. Plans were also established for the 2008 Common Computing Readiness Challenge, which will have all four experiments running simultaneous exercises.
Jamie Shiers ran the workshop and reported, "It was very productive in terms of understanding where we are with the overall service and what remains to be done for final service commissioning over the coming months. The biggest problem is almost certainly the ramp-up in terms of scale: capacity, performance and reliability. It was very encouraging to see such a large number of people sufficiently devoted to the service that - once again - they were prepared to give up their weekend".
On Monday, the conference itself opened with an update on progress of the LHC machine and its experiments by Jim Virdee of CERN and Imperial College, who presented the prospects for physics from the LHC in 2008. This was followed by Les Robertson, Leader of the LCG project, looking at the challenges when the data begins to flow, showing the massive ramp up required in Tier-1 and Tier-2 CPU and the increase in reliability needed if computing is to help the fast delivery of physics results. During the parallel sessions, attendees were split between eight sessions, with GridPP members presenting in five of these:
- Software components, tools and databases included a talk from Caitriana Nicholson of Glasgow, on the ATLAS Tag database.
- Computer facilities, production grids and networking had two contributions from the UK, with a talk by GridPP's Production Manager Jeremy Coles on running the production Grid, including findings from recent infrastructure and service challenges as well as a look forward to how ready GridPP is for the LHC. The session also saw a presentation from Alexandra Forti about storage systems, comparing at dcache, xrootd and slashgrid.
- the Grid middleware and tools track showed the best representation from GridPP, with talks by Steve Fisher on R-GMA, Stephen Burke about the GLUE information schema, Andrew McNab presenting GridSite security and Greig Cowan talking about storage at Tier-2s.
- UK developments were well represented in the Distributed data analysis and information systems track. Seven talks covered the use of the GANGA framework by ATLAS and LHCb and the DIRAC workload management system in LHCb. The CMS CRAB distributed analysis system was also presented, as was the ASAP testbed for gLite. Roger Jones chaired the session, and concluded, "It's good to see that each of the LHC experiments now has a near-live distributed analysis framework, and the running experiments are starting to benefit from this too. But there's still a lot of scope for re-use of tools already developed, and for common projects between experiments." He also noted the huge challenge posed by efficient data access, with interesting UK work on data selection and access using database tags presented by Helen McGlone.
- Collaborative tools and initiatives was a shorter session, chaired by Pete Clarke and including a talk on EVO (Enabling Virtual Organizations), the latest videoconferencing technology from Caltech and used extensively in GridPP. This session also saw one of the non -GridPP UK talks by Prof. Richard McClatchey of the University of the West of England, on bringing together pediatric information for the Health e-Child project, which has Great Ormond Street as a partner.
CHEP also played host to an exhibition, at which GridPP were well represented showing the Real Time Monitor, an example of Imense's image searching technology (works very well so far with pink flowers on a green background) and a well-received demonstration by Ying Ying Li of GANGA on Windows. Neasan O'Neill, GridPP's Events Officer, said, "This was definitely one of our better conferences in terms of interest in the stand. And it's still surprising to find how many particle physicists don't know what GridPP is or how our Grid works." GridPP had several posters in the two poster sessions, including one from Graeme Stewart on LAN access to storage elements and a view from the Tier-1 by Martin Bly.
For many the highlight of the trip was an opportunity to go whale watching on the free afternoon. Although missed by most of the GridPP contingent for various spurious reasons (seasickness, a security workshop, 'writing my talk'), those who did go were universally enthusiastic despite a touch of sunstroke, with every boat seeing orcas. The conference dinner was also a great success, held in the Museum of British Columbia with food distributed around various galleries, giving the chance to eat dessert in with the totem poles or drink surrounded by stuffed native Canadian wildlife.
More photos from the conference are at: http://www.gridpp.ac.uk/chep07/index.html
© Copyright GridPP
If you wish to reproduce this piece please credit GridPP and contact Neasan O'Neill to say you are using it