As I pointed out a few weeks ago, cloud computing is not always the right tool for the job -- specifically, when there are high-end computational requirements and the processors need to be more tightly coupled to support high-performance computing standards such as MPI.
However, infrastructure service providers, such as Amazon Web Services, are perhaps finally putting the right tools in the shed. That seems to be Amazon.com's intent with its new high-performance-computing (HPC) cloud service offering, which could provide a home for those special applications that require supercomputing power.
[ Get the no-nonsense explanations and advice you need to take real advantage of cloud computing in the InfoWorld editors' 21-page Cloud Computing Deep Dive PDF special report, featuring an exclusive excerpt from David Linthicum's new book on cloud architecture. | Stay up on the cloud with InfoWorld's Cloud Computing Report newsletter. ]
Amazon.com's Cluster Compute Quadruple Extra Large is a 64-bit platform with 23GB of memory, 1,690GB of instance storage, and 10Gbps of Ethernet I/O performance composed of 33.5 Elastic Compute Cloud (EC2) compute units. The default usage limit for the Cluster Compute instance type is eight instances, or 64 cores, although customers can request more, the company says.
That's hardly the kind of genetic-workload infrastructure platform Amazon Web Services is best known for. Even critics are warming up to the upmarket offering. For example, although the National Energy Research Scientific Computing Center (NERSC) at the Lawrence Berkeley National Laborator found Amazon.com's EC2 wanting, NERSC has now determined that its HPC applications run 8.5 times faster than on previous EC2 instance types, according to Amazon.com.
The issue with EC2 does not concern processor power as it does the network bandwidth between those processors. High-performance cluster computing needs huge pipes between processor instances, and most infrastructure cloud providers don't make the grade. Amazon.com's Cluster Compute appears to come through. Thus, Amazon.com claims that "the service is appropriate for tightly coupled parallel processes or applications sensitive to network performance that use node-to-node communication, typically high-performance computing applications."
This is a great move by Amazon Web Services, especially considering it is now offering computational power that most enterprises don't have in their own data centers due to cost or specialization. And Amazon.com is charging just $1.60 per hour, which will make it even harder for IT to justify investments in high-end equipment for those special cases that need this type of high-performance computing.
Once again, Amazon.com sets a standard that other cloud providers will surely follow.
This article, "Amazon.com's big move into high-end computing," originally appeared at InfoWorld.com. Read more of David Linthicum's Cloud Computing blog and follow the latest developments in cloud computing at InfoWorld.com.