For $1.60 per hour, the Cluster Compute Instances, or cc1.4xlarge, offer:
- 23 GB of memory
- 33.5 EC2 Compute Units (2 x Intel Xeon X5570, quad-core "Nehalem" architecture)
- 1690 GB of instance storage
- 64-bit platform
- 10 Gigabit Ethernet
It won't even be AWS' most expensive instance, which is the High-Memory Quadruple Extra-Large at $2.40 per hour.
Amazon users already in awe of new instance
Reaction has been swift. Rob Gillen, a researcher and developer at Oak Ridge National Laboratories in Tennessee, said colleagues have already calculated what the new instance could mean to them.
"One of the computational biologists here just noted that his department could replace their cluster with an [AWS] cluster compute for $208 per hour," he said on Twitter.
Amazon says that the new instances differ from standard AWS offerings in more ways than one. First, ccc1.4xlarges run on a cluster that is all new, top-of-the-line hardware built to operate like a traditional grid computing environment.
It's only to be found in Amazon's Ashburn, Va. data center, and Amazon says users will get "full bisection 10 Gbps between instances" if they run more than one instance. They are also disclosing the CPU model so HPC users can tune their Linux cluster operating systems appropriately, something you don't get with regular Amazon EC2.
Supercomputer users should not get greedy just yet, however, as AWS limits you to eight cc1.4xlarges at a time. Still, HPC use cases on Amazon have already been documented. The Relativistic Heavy Ion Collider (RHIC) in Brookhaven, New York demonstrated the ability of EC2 to run clusters effectively last year. In fact, computer scientist Kate Keahy said her colleagues' favorite after-dinner pastime was to pull out napkins and calculate their cluster cost on Amazon.
Carl Brooks is the Technology Writer at SearchCloudComputing.com. Contact him at firstname.lastname@example.org.
Dig deeper on Amazon EC2 (Elastic Compute Cloud) management