Community Cluster Program

Overview

The Purdue Rosen Center for Advanced Computing operates a significant cluster computing infrastructure, which was developed over several years though focused acquisitions using funds from grants, faculty startup packages, and institutional sources. The Community Cluster Program (CCP) has been a great success, and over $1.2M was invested in computational and storage resources in Q4 2006 and Q1 2007.  The CCP is now the foundation for Purdue's research cyberinfrastructure and will have a significant role as RCAC plans for the retirement of aging HPC resources such as the IBM SP and Sun F6800 cluster.

RCAC is expanding our community cluster options with Phase II of the Community Cluster Program. We are offering a subscription service and a ready-to-rack option beginning in early 2008.

For more information please send email to: rcac-info@purdue.edu

 

Subscription Service:

  • ITaP owned continually updated cluster.
  • You subscribe to the number of cores you need and you are computing within a week.
  • You set the time frame for your subscription from months to 1 year, 2 years, or 3 years.
  • We provide stable, secure, up-to-date computing resources for you and your research group.

Ready-to-Rack:

  • You purchase machines from a menu of tested configurations and are up and running in about one month.
  • We provide stable, secure, up-to-date computing resources for you and your research group.

Community Cluster Phase II options provide the research faculty with the following benefits:

  • Peace of Mind: RCAC system administrators take care of security patches, attempted hacks, operating system upgrades, and hardware repair so faculty and graduate students can concentrate on research.
  • RCAC data center provides infrastructure such as networking, racks, floor space, cooling and power.
  • Cost effective: RCAC works with vendors to obtain the best price for the computing resources while combining financial resources from disparate sources to provide additional group purchasing power.
  • Much quicker time frame (approximately 1 week to one month) to be up and computing as compared to a custom cluster configuration which usually takes about 6 months from start to production.

Frequently Asked Questions

Last Updated 12/06/07