Seven Ways to Make Big Data an Actionable Opportunity
Normally storage companies love the idea that Big Data is going to result in consumption of more storage, but when the sheer volume of data involved starts to exceed the available amount of customer funds available, another approach must be found.
With that issue in mind, EMC and RainStor, a provider of database software, have inked a partnership under which they will bundle RainStor’s software with Isilon network attached storage (NAS) systems running the Hadoop Distributed File System (HDFS).
Deirdre Mahon, vice president of marketing for RainStor, says what makes RainStor unique is that while it compresses data by as much as a factor of 40, that information can still be queried via a MapReduce or SQL interface. So instead of archiving massive amounts of data to tape and then having to reload it whenever a Big Data application needs that data, RainStor allows data to be efficiently stored on, for example, an Isilon scale-out NAS system or a write-once, read many (WORM) optical disk deployed on another storage system while still being considered active for an application.
The ultimate Big Data goal is to create petabytes of data that organizations can readily mine to discover patterns and trends. While Hadoop provides a comparatively inexpensive way to manage massive amounts of data, the larger a Hadoop cluster gets, the more difficult it is to manage. Deployed alongside Hadoop, Mahon says RainStor provides an efficient way to keep massive amounts of data active without having to invest in additional servers and storage systems every time the amount of data in the Hadoop cluster grows.
Obviously, not every IT infrastructure vendor is going to be in love with the idea of selling less IT infrastructure hardware. But given the fact that the total cost of Big Data encompasses much more than the hardware involved, EMC in partnership with RainStor is freeing up some dollars that could be better applied to developing an actual Big Data application rather than simply housing massive amounts of data that may only be needed infrequently.
