A publication of the Office of Advanced Simulation & Computing, NA-114, NNSA Defense Programs
December 2007
NA-ASC-500-07—Issue
Return to this issue’s stories
High-Performance Computer Support Team Reinvents Itself
“Better, Faster, Cheaper” is the three-pronged goal of Los Alamos National Laboratory’s High-Performance Computer Support Team (HPCST) for in-house computer hardware support. The HPCST has reinvented itself by redefining its long-established mission of monitoring all LANL’s supercomputers, clusters, and facilities. In addition to monitoring these systems, the team has taken on the responsibility of providing maintenance and support for all the supercomputer systems currently installed at the laboratory. The HPCST staffs the operations center in the Metropolis Center for Modeling & Simulation 24 hours a day, 7 days a week. It was this availability of personnel, in part, that sparked the innovative initiative to provide in-house computer maintenance and support services.
The benefits realized by the hardware support initiative have gone beyond the reduction of maintenance costs typically paid to a vendor support company. The Mean Time to Repair (MTTR) figures are also significantly reduced across all systems because the response time to hardware failures has been virtually eliminated. The lower MTTR has resulted in higher system availability figures.
The team has been able to accomplish this goal by making some strategic hires and providing computer hardware repair training for all of its members. Every member of the team can quickly diagnose and isolate a failed node in a system and replace it to bring the cluster up to 100% availability. The failed nodes are then repaired and tested offline utilizing a testing workstation that the team also maintains.
The maintenance and repair of supercomputer systems at LANL has provided a greater technical challenge for the HPCST staff. The computer operator’s job has evolved from hands-off monitoring to monitoring and a hands-on maintenance and repair effort. Technical challenge, cost savings, and success have resulted in higher job satisfaction for the members of LANL’s HPCST.
DOE Privacy Disclaimer | Sandia Privacy Disclaimer | SAND 2007-8167 W
Developed and maintained by Sandia National Laboratories for NA.114