On Monday, Amazon Web Services (AWS) announced the Beta launch of Cluster Compute Eight Extra Large (CC2), a supercomputer on its EC2 cloud infrastructure that ranks 42 on the list of the world’s top 500 supercomputers. Cluster Compute Eight Extra Large can be used for “physics simulations, seismic analysis, drug design, genome analysis, aircraft design and a variety of business computing and analytics applications.” In addition, the high performance computing (HPC) application can be used with Amazon Elastic MapReduce to analyze massive amounts of data from structured and unstructured data sets using the Hadoop software framework. Amazon Web Services revealed the deployment of Cluster Compute Eight Extra Large in conjunction with this year’s November 12-18 Supercomputing 11 conference in Seattle. The announcement represents a clear strategic move to increase enterprise market share by targeting organizations such as pharmaceutical manufacturers, large scientific research groups and financial services firms that encounter the challenge of crunching massive amounts of data.
Key features of Cluster Compute Eight Extra Large include the following:
• 2 Intel Xeon processors, each with 8 hardware cores per instance
• Hyper-threading that allows each core to process instructions in parallel
• A whopping 60.5 GB of RAM and 3.37 TB of instance storage
• Choice of Amazon Linux AMI or Windows 2008 R2 for the operating system
• Availability only in Amazon’s US East (N. Virginia) Region
• Price of $2.40 per hour per instance
A cluster of 1064 CC2 instances resulted in a speed of 240.09 teraFLOPs. 290 CC2 instances resulted in a speed of 63.07 teraFLOPs with pricing less than $1000/hour. The Cluster Compute Eight Extra Large product represents an upgrade to High Performance Computing offerings introduced by Amazon Web Services last year. Customers of its HPC offering currently include Harvard Medical School, which uses Amazon’s HPC product to investigate problems in genomics and personalized medicine. Lawrence Berkeley National Labs was another user of Amazon’s initial HPC offering known as Cluster Compute Instances.
The release of CC2 positions Amazon Web Services strongly in a market dominated by the likes of IBM, Cray, HP and Dell. Because supercomputers often start in the price range of $500,000 and up, AWS seeks to compete in price and ease of deployment. The offering also represents yet another tactic to gain market share in the hotly contested Big Data space, particularly given the compatibility of CC2 with Amazon Elastic MapReduce and its Hadoop based infrastructure. In the cloud computing space, however, Cluster Compute Eight Extra Large represents a clear move to consolidate market share in the enterprise space by democratizing supercomputer usage beyond the domain of scientific and research organizations.