Continuing to extend its high-performance computing (HPC) expertise into the realm of Big Data, SGI announced today new Big Data storage offerings that not only include a Cloudera Hadoop distribution running on Red Hat Linux, but also provide a one-to-one ratio between server cores and disk spindles.
In addition, SGI announced it would make available a cloud-based version of this offering for organizations looking to build Big Data proof-of-concept applications.
While SGI doesn’t think production Big Data applications are cost-effective to run in the cloud, Bob Braham, senior vice president and chief marketing officer at SGI, says the cloud offering is designed to give organizations a sandbox where they can get started with Hadoop without having to make a massive upfront investment.
The SGI InfiniteData Cluster makes use of Intel Xeon processors E5-2600 v2 and up to 12 4TB drives per tray to provide support for up to 40 compute nodes capable of supporting 1.9PB in a single rack. The Hadoop distribution from Cloudera comes pre-integrated.
SGI today also announced that it has developed SGI ObjectStore, an object-based storage system that makes it easier to scale out Big Data applications and SGI InfiniteStorage Gateway, an archiving appliance.
In the wake of SGI’s recent acquisition of FileTek, Braham says SGI is gearing up to establish a much more significant presence in the enterprise. While well known in HPC circles, Braham says the rise of Big Data in the enterprise is creating storage challenges that SGI has already solved in HPC environments. The challenge now, says Braham, is bringing that technology to bear in a way that is easily consumed by enterprise IT staffs that are not all that familiar with SGI.
The degree to which SGI can do that in an already crowded enterprise IT storage field remains to be seen. But as IT organizations begin to struggle with Big Data, chances are that more than a few of them are going to be more than willing to consider additional options.