Amazon has already gained quite a customer base for its cloud-based EC2 computing services, but now it's ready to pull in bigger customers. A new program, called Cluster Computer, can deliver a level of supercomputing power most typically seen in research institutions and universities. In one test, an "880-node cluster attained 41.82 teraflops, or floating-point operations per second, using the Linpack mathematical speed test--that's pretty close to the performance of the 145th fastest supercomputer in the world." Why bother? According to an Amazon spokesman:

EC2 is now used for workloads ranging from genomic sequence analysis to financial modeling and automotive design. The Cluster Computer Instances are designed to support high performance computing tasks, including parallel processing workloads. 'These customers have told us that many of their largest, most complex workloads required additional network performance, (the spokesman) said.

Access to the Cluster server $1.60 per hour, about seven times more than access to regular small EC2 servers. You can read much more about the speeds and feeds as you ponder the interesting question of what you could use all this raw computing power for. Typical businesses don't need supercomputing finesse, and those that do may be wary of using a cloud-based solution, but if anyone can pull it off, Amazon can. EC2 has been a roaring success, and this is the next logical step. -- Don Willmott