The CMS Computing System: Successes and Challenges

Each LHC experiment will produce datasets with sizes of order one petabyte per year. All of this data must be stored, processed, transferred, simulated and analyzed, which requires a computing system of a larger scale than ever mounted for any particle physics experiment, and possibly for any enterprise in the world. I discuss how CMS has chosen to address these challenges, focusing on recent tests of the system that demonstrate the experiment's readiness for producing physics results with the first LHC data.