Sarah Pearce (CASS/ATNF)

Grid computing for the LHC - can astronomers learn anything from particle physicists? - Sarah Pearce Colloquium

ATNF Marsfield Lecture Theatre

Abstract

CERN's Large Hadron Collider produces up to 15 petabytes of data per year, for long term storage and distribution to thousands of physicists around the globe. To deal with analysing this data deluge, the LHC has constructed the world's largest computing grid, consisting of 200,000 CPUs spread at nearly 600 sites worldwide. As well as running upwards of 1 million computing jobs per day, the grid stores petabytes of data on disk and tape. This talk will examine the LHC grid, looking its structure, how the software manages data and processing, and issues such as security. I will also define and discuss other types of distributed computing, such as cloud and volunteer computing. ASKAP will have similar data rates to the LHC and, as a survey telescope, will have to deal with providing data access for large numbers of users. Is there anything astronomy can learn about computing from the LHC experience?