Supporting supercomputer users
Computer Production Group (CPG) staff,
represented by Raisa Leifer, Scott Baker, and Susan Albertson (left to
right), provide 24x7x365 frontline customer support and computer room
operations and monitoring. CPG staff, always available to handle any
contingency, are critical for the business continuity of the computing
facility and the institution.
NCAR's strategic commitment is to provide robust, accessible, and innovative
information services and tools to our customers. CISL provides end-to-end
service for its supercomputing users using the full capability of the
laboratory staff. For increased efficiency, consulting service has been
expanded to include other groups within CISL who provide frontline
assistance as well as in-depth expertise. Contacts are tracked using an
ExtraView HelpDesk trouble ticket system.
In FY 2008, frontline support resolved 1,582 tickets in the time interval from September 2007 through August 2008. In the same interval, advanced support resolved an additional 1,061 contacts, for a total of 2,643. The average number of log entries per ticket was 4.5, with communication highest with users on complex cases. Average response time for frontline support for ticket resolution was about 1.3 days, while a longer average response time of about 4.2 days was required for more complex issues. The average number of staff who worked on tickets was 1.31, demonstrating cross-team cooperation in resolving issues.
A benefit of tiered customer support has been to free CISL staff to supply customized, one-on-one service for special campaigns such as the Accelerated Scientific Discovery program, which is providing opportunities for capability computing projects requiring hundreds of thousands of processor hours on the new IBM POWER6 platform, bluefire. As the system was brought online, the consulting group provided input on making the system easier to access, documentation and training on the new features of the system, and help to users in porting and scaling their parallel codes to run well on the new platform.
In FY2009, we anticipate further growth in this type of scientific support, as well as to continue support for new versions of NCAR "flagship" models such as CCSM and WRF. We also expect to increase support for TeraGrid-allocated projects.
This work supports NCAR's strategic goal to "Provide robust, accessible, and innovative information services and tools" to our customers. This ongoing service is supported by NSF Core funds including CSL funding.