This slidecast from Intel’s Tom Stachura provides a nice overview of network requirements for doing HPC in the Cloud.
At IDF 2010 in SF, we had the opportunity to demonstrate a nice use case for HPC in the Cloud – when you have tapped out your local cluster resources, provision your excess work to the Cloud. 10GbE iWARP was used for the local cluster as the performant low-latency fabric. Mainstream 10GbE was used in the Cloud as it provides the dynamic virtualization and unified networking features required for virtual data centers. See the class I presented on HPC in the Cloud networking – it covers some additional usage models and some key features required for cloud networking.”