Univa UD announced the publication of a white paper and supporting tools that documents the process of implementing its open source UniCluster product in Amazon’s Elastic Computing Cloud (EC2) web service.

Univa UD sponsored BioTeam, a life sciences informatics consultancy and systems integrator, to document the process and publish the tools. The paper titled How to: UniCluster and Amazon EC2, presents step by step instructions for setting up UniCluster within EC2. UniCluster is an open source cluster software stack, integrating widely used open source components including Sun Grid Engine, into a free, downloadable product that simplifies cluster deployment, configuration and management. UniCluster is currently available for download from Grid.org (www.grid.org), the online community for open source cluster and grid software users, administrators and developers. Optional Annual Care cluster enterprise-level support is also available.

Amazon EC2 is a web service that provides resizable compute capacity in order to make web-scale computing easier for developers. EC2 presents a true virtual computing environment, providing users with web service interfaces to requisition machines for use, load them with a custom application environment, manage network access permissions, and run an image using as many or few systems as desired.

“This approach matches how we have seen some of our more advanced clients and partners handle WAN-scale computing needs,” said Chris Dagdigian, Principal Consultant of BioTeam Inc., and author of the paper. “Often there are core, company operated systems handling authentication, data staging and workflow while the remote systems operating across Metro or WAN-scale networking distances are merely used as engines for computation.”

The white paper provides a summary of a UniCluster deployment within the EC2 cloud giving enough detail to allow others to replicate the work. Also included in the paper are workarounds to any technical issues encountered, a public AMI image with UniCluster and associated deployment and configuration scripts.

"In recent years we have seen increasing adoption of HPC technologies in business areas outside the traditional boundaries of science and engineering analysis," said Addison Snell, Vice President of Tabor Research.

"Big-data applications and cloud computing deployments increasingly rely on scalable technology that comes from vendors with HPC expertise."

“A growing number of customers are looking to virtualization as a means to ‘buy to baseline; provision to peak’ for their high performance compute workload,” said Gary Tyreman, Vice President and General Manager of HPC for Univa UD. “The best practices published as a result of this project, extend Univa UD’s cluster software management capability securely to the cloud, a key step in the operational aspect of the HPC lifecycle management. This uniquely enables our customers to readjust their available compute power on the fly, which will translate to tremendous operating expense savings.”

The white paper is available for download at www.univaud.com.