Tiers of Joy: The new RAL computing centre

The day started around 2pm with Neil Geddes, the UK representative on the EGI council welcoming everyone and explaining the drive for the grid from the LHC. This was just as the grid (and the Tier 1) was starting to see the first effects of the 7TeV collisions at CERN with data transfer rates soaring. After Neil were talks from the teams who provide the services in RAL's computing group; hardware management, grid services, large scale data storage, databases and the production team who hold it all together. Throughout the talks the tier 1 team were also running tours of R89 which were all very busy and highly informative.

Once the technical talks and tours had ended it was time for the official opening. This included talks from Keith Mason the STFC chief executive, Andrew Taylor, the head of ISIS, and David Britton, GridPP's project leader, as well as some wine and nibbles to celebrate the new facility but also toast the start of a new era for particle physics and the grid.

The planning for the new machine room at RAL started in 2005 but it wasn't till August 2007 that building actually commenced. With 892 square metres of machine room space, 80 racks of computers, a tape robot, four 750kW chillers, four 2350kW transformers and a 120kW UPS for critical services (backed up with a diesel generator) it was not a small undertaking. By May 2009 however most of the work was done and staff were able to move into their new shiny offices. However the computing kit was another matter. The crew at RAL were well aware of the difficulties involved in moving a facility of this magnitude and had been planning the move for months beforehand. The move began only 3 days after the end of STEP'09 but in less than 3 weeks on the 6th of July the team could sit back and declare the UK's Tier 1 back up and running.

As the UK's Tier 1 site for LHC computing RAL is one of only 11 Tier1s in the world spread across the 10 countries; Canada, France, Germany, Italy, the Netherlands, the Nordic countries, Spain, Taipei, the UK, and two sites in the USA. Moving such a system has only been attempted a few times within the grid community and any major downtime caused by hiccups would effect not only the UK but hinder global operations.

The full story of the move from the old computing centre is told in the Tier 1 blog here: http://www.gridpp.rl.ac.uk/blog/category/r89-migration/