Support of Supercomputer Users

This schematic of CISL's three-tier Customer Support shows how the tiers relate to the supercomputer (high-end) customer. Customer Support is integrated throughout the entire CISL organization. Well-managed interaction between the frontline (Tier 1), Consulting Services (Tier 2), and supercomputer administrators (Tier 3) provides a full spectrum of support targeted to supercomputer user needs while optimizing CISL staff's commitment of time and expertise.

NCAR has provided their supercomputer users with expert assistance since the inception of the NCAR computing center. This year, CISL staff supported users across all of the production supercomputer clusters via the ExtraView HelpDesk trouble ticket system that began operating on March 5, 2006. This support is organized into three tiers including a 24 by 7 frontline, a software engineering consulting tier, and an advanced support tier. This support is a regular feature of the services CISL provides to the scientific user community, spanning a range of issues from frequently asked questions with simple answers to complex problems requiring several weeks of specialized assistance. This ongoing service maintains NCAR's strategic priority of "Developing and providing advanced services and tools."

In FY 2006, Tier 1 frontline support resolved 613 tickets in the time interval from March 14 to August 18, 2006. In the same interval, Tier 2 resolved 306 tickets. Tier 1 offers 24 by 7 customer support, answering many questions directly, and submitting tickets on behalf of customers for further assistance by Tiers 2 and 3. Customers and all CISL Support Tiers benefit from this arrangement, since the tickets are archived and searchable for future problem resolution, and since the customer support is consistent across phones, email, and the web-based trouble ticket system.

In FY 2007, CISL staff will continue to provide three-tiered support for our supercomputer customers, adding new customers who arrive via the TeraGrid. Software engineering work will continue to assist customers in all aspects of their supercomputing workflow including parallel programming, debugging, code optimization, compilers and libraries, postprocessing applications, migrating jobs to run under the Load Sharing Facility (LSF) batch system, and Special Campaigns.

This ongoing service is supported by NSF Core funds including CSL funding.