That means Windows-centric customers can now use Amazon’s “compute intensive beasts” to run heavy-duty media handling, rendering or computational finance work loads, according to an Amazon Web Services blog post. That pretty much translates to boosting Amazon’s enterprise cachet, as enterprises seek ways to analyze and crunch large amounts of data more cheaply.
The extra-large compute cluster weighs in at 23 GB of memory; 33.5 EC2 Compute units (two Intel Xeon X5570, quad-core “Nehalem” processors); and 1690 GB of storage. The Cluster GPU sports pretty much the same specs but adds support for two NVIDIA Tesla “Fermi” M2050 graphic processing units.
Microsoft Windows Server 2008 R2 loads could already run on AWS’s smaller compute clusters but this move gives Windows more parity across cluster types with Linux. Before this week, customers could select Amazon Linux AMI, and SUSE Linux from the AWS Console Quickstart although an Amazon spokeswoman said customers are also running Ubuntu, CentOS and FreeBSD on Cluster Instances as well.
These kinds of big compute clusters are ideal for converting lots of graphical data or running intense quantitative analysis, but may not be that interesting to many rank-and-file businesses, said one long-time AWS user. Still, the news is noteworthy in that it shows that Amazon is trying to be open handed in its operating system support.
AWS is trying to embrace and cooperate with Microsoft and other ISV incumbents from the on-premises software world. In the past there’s been tension between AWS’s utility-based price model and the per-machine licensing that reigns in the on-premises computing world.
This spring, Amazon Web Services started pairing its Spot Instances and Cluster Compute Instances, a combo that appealed to businesses with ad hoc batch-processing needs. AWS started offering high performance computing (HPC) through its compute cluster instances in April 2010 and added GPU Instances last November.