Blog Post

Amazon supersizes compute cluster

Amazon (s AMZN) Web Services on Tuesday beefed up its Cluster Compute family with an extra-large iteration that packs in two eight-core Intel Xeon processors each connected to a 10-Gigabit network.

With this option, Amazon continues to make bite-sized chunks of its cloud services available to users of all sizes and computing needs. This is just the latest — and biggest instance — and brings the total of Amazon instance types to 12, which vary according to CPU, networking, storage and memory.

Cluster Compute Eight Extra Large (dubbed cc2.8xlarge of CC2 in AWS parlance), available now in a beta, comes with 60.5 GB of RAM and 3.37 TB of instance storage, according to the Amazon Web Services blog.

The new compute cluster should appeal to high-performance computing (HPC) enthusiasts who want to offload some or all of their big jobs to Amazon’s cloud.  Users can opt to run CC2 instances on Linux Amazon Machine Images (AMIs) or Windows Server 2008 R2.

Developers can launch an on-demand CC2 instance for $2.40 per hour for Linux of $2.97 per hour for Windows.

The previous largest cluster (the Quadruple Extra Large cluster) started at $1.30 per hour for Linux and $1.61 per hour for Windows.

Users can also reserve instances for range of prices. Or they can bid to run CC2 on the spot market as needed. Use of such spot instances is attractive to companies that need to run batch-processing jobs on an ad hoc basis. This may all seem like penny ante stuff, but Amazon has built AWS into a billion-dollar-a-year business over the past decade.

Amazon has opened more options since launching its first HPC instances in July 2010. This summer, it started pairing spot instance pricing with HPC instances. In July 2010, it launched HPC instances for the first time.  The HPC effort has made some waves; Amazon now stakes its claim in the list of  Top500 supercomputer sites, where it is ranked number 42.

Since this is a beta release there are some limitations — CC2 is available in a single zone in the northeast U.S. region with capacity to be added to other regions throughout next year.

Photo courtesy of Flickr user turtlemom4bacon

4 Responses to “Amazon supersizes compute cluster”

  1. Amazon is spreading their paws in enterprises world and providing their best services in business world.Its good that they have added something new in their cluster and i am sure that they will definitely bang the IT world

  2. Garrett Nix

    Typo: “The new compute cluster should appeal to high-performance computing (HPC) enthusiasts who want to offload come or all of their big jobs to Amazon’s cloud.”

    I think you probably mean, “offload some.”