CERN

Overview
CERN needed a way to manage the growing amount of data collected when sub- atomic particles are smashed together in high-energy accelerators, and the increasing application workload generated by jobs submitted by 6,500 scientists worldwide.

Challenges
CERN needed to harness the full potential of its computing resources in order to optimize the performance of compute- and data-intensive applications, while reducing support requirements and creating a reliable computing environment that will scale easily as requirements grow drastically over the next few years.

Solution

• Platform LSF Software

Results

• Existing computing cluster is achieving new levels of application performance, and results are being delivered to users when they need them

• Minimal support or manual intervention is required to keep workload management system running smoothly, thus reducing staff requirements

• Complex data management software workload is also being managed, enabling different users to access many data sets as needed

• The application processing and data management environments are fault tolerant, unaffected by periodic individual CPU failures

• Platform LSF workload management solution is successfully scaling up as computing infrastructure and workload grow

"We've had a very fruitful collaboration with Platform. We've both benefited from this in terms of improvements in areas such as queuing techniques and failover, and in the interesting new ways of using the Platform LSF product, such as for grid-level scheduling."

Dr. Tony Cass
Leader of CERN's Computing Fabric, Infrastructure & Operations Group