The Challenge of Keeping Up With the Data

The immense amount of data emerging from the Relativistic Heavy Ion Collider (RHIC) gives the staff at the RHIC/ATLAS Computing Facility (RCF/ACF) plenty of work to do. In five years of running, the four RHIC experiments have taken digitized “snapshots” of billions of particle interactions – data-dense “pictures” that may reveal details about the early structure of the universe and the fundamental properties of matter.

During RHIC’s 2005 run, for example, the experiments collected a total of 675 terabytes of data — enough to fill roughly one million compact discs. As the ions were colliding, the RCF received data from the experiments at rates in excess of 200 megabytes (MB) per second — equivalent to transferring the contents of a CD in a mere three seconds. All those terabytes were stored on robotically manipulated tapes for later analysis.

The 2006 RHIC run is now complete. The data is being stored in a new robotic tape storage system which has the theoretical capability of recording data at rates in excess of 1000 MB per second and has a data storage capacity of over 2.4 petabytes.

Now, the 3,300 processors making up the RCF’s RHIC Linux Farm are crunching the numbers. Year-round, the RCF/ACF staff are working 24 hours a day, seven days a week, providing both hardware and software support for users. “When needed, they repair, replace, or upgrade facility hardware and deploy and upgrade system software,” said Bruce Gibbard of the Physics Department, who heads the facility. “The staff also uses sophisticated software to monitor computer system usage, performance, and status, and make sure data back-up, e-mail, and web services are working smoothly.”