SCIENCE

Clustercorp has announced general availability of Rocks+ for Amazon Cluster Compute Instances and Cluster GPU Instances. Organizations with large or variable computational demands can leverage Rocks+ to power massively scalable clusters in the cloud, while benefiting from the elasticity, flexibility and cost advantages of Amazon EC2.

Rocks+ dramatically reduces the time and cost of setting up and managing Cluster Compute and Cluster GPU solutions in the cloud through intelligent software automation. The entire software stack is packaged and deployed as a monolithic, yet modular Linux distribution. This system provides for a single step install, which is capable of dynamically provisioning heterogeneous appliance types at massive scale (in parallel, leveraging BitTorrent-style package sharing). A key differentiator that separates Rocks+ from other "complete stack" paradigms, is a modular framework called Rolls, which are optional, automatically configured, multi-instance software systems connected by the Rocks Graph. Rolls allow customers to take advantage of an end-to-end, packaged stack and still choose which software components are included to meet site-specific requirements.

"Cluster GPU and Cluster Compute Instances provides our customer with highly available, scalable and affordable resources for high-performance computing (HPC) in the cloud," said Peter De Santis, general manager, Amazon EC2. "We're excited to work with Clustercorp to help our customers access its familiar management tools to run a variety of HPC application workloads, ranging from marketing and financial analytics to automotive and aerospace design."

Rocks+ is available to end-users direct from Clustercorp, through Amazon Web Services as an AMI, and through multiple hardware partners including H-P and Dell.

"Clustercorp is thrilled to be launching Rocks+ for more Amazon EC2 instance types," said Tim McIntire, CEO and co-founder of Clustercorp. "We are committed to providing cloud administrators the same powerful tools that have become the de facto standard in the cluster computing space."