Comments on: Supercomputer vet Cray wants to turn big data into fast data http://gigaom.com/2012/02/08/supercomputer-vet-cray-wants-to-turn-big-data-into-fast-data/ Tue, 21 May 2013 11:52:50 +0000 hourly 1 http://wordpress.com/ By: Bill McColl http://gigaom.com/2012/02/08/supercomputer-vet-cray-wants-to-turn-big-data-into-fast-data/#comment-807726 Thu, 09 Feb 2012 23:11:07 +0000 http://gigaom.com/?p=482281#comment-807726 Interesting. Cloudscale’s patented HadoopBI big data engine (hadoopbi.com) has been designed from the ground up to analyze streaming data in realtime at rates of millions of events per second. Like Cray we use MPI (HadoopMPI) to achieve performance that is over 100x faster than HadoopMapReduce. HadoopBI can be run on Apache HDFS or on MapR, but for blazingly fast performance it can also be run on Lustre, which is again is way faster at scale than any of the current Hadoop filesystems. Lustre is also what Cray use in their supercomputers. HadoopBI runs on standard commodity hardware.

]]>