As enterprise adoption of Hadoop increases, capabilities around ease-of-use, efficient scale and reliability become a high priority. To date, Hadoop has been deployed on low-cost commodity direct attached storage (DAS) environments but different architectures are now being deployed to optimize both compute and storage. As data volumes grow, the need to balance CPU and storage capacity is critical in order to achieve an optimal environment. With DAS deployments, compute and storage are rigid and cannot be separated which becomes challenging as requirements change over time.
RainStor enables enterprise customers to run Hadoop anywhere. With unique data compression yielding a 20-40X storage reduction, companies can scale their data with greater predictability and efficiency. The RainStor highly-compressed file format is virtualized from the underlying storage layer and therefore behaves the same whether running on DAS or NAS. With the ability to co-exist Hadoop data running on both DAS and NAS, you can now separate the compute layer from the storage layer and gain both scale efficiencies and query performance.
"Isilon running on RainStor provides high-impact use cases, including a compliance data archive for years of history reaching petabyte scale", stated Sam Grocott, Vice President, Marketing and Product Management, EMC Isilon Storage Division. "The one-two punch of RainStor and Hadoop on Isilon gives customers both performance and efficient scale with the added bonus of being easy to deploy and maintain. Case in point: RainStor and Isilon customer saw a 32x compression rate enabling efficient, predictable scale."
Running Hadoop on Isilon with RainStor, you achieve:
"RainStor has allowed AdaptiveMobile to support a large telecommunications customer for a number of years managing over two petabytes of communications data loading around 1 billion records daily", stated Gareth MacLachlan, chief operating officer, AdaptiveMobile, a long-time RainStor partner. "We believe RainStor on scale-out Isilon NAS is a highly efficient way to manage petabyte volume, with excellent performant load and query capabilities", added Gareth MacLachlan.
"With RainStor's compression and efficiency, we enable new Big Data architectures that were simply not possible before", stated John Bantleman, CEO of RainStor. "EMC's Isilon scale-out NAS is ideal for enterprises that have skills in-house that know how to manage these environments. Our customers are looking for new architectures where they can combine a one-petabyte Hadoop cluster with a multi-petabyte NAS store for mass storage, no compromise for query and the lowest TCO."
"With RainStor the network bottleneck is eliminated when moving data from storage to compute because of the high level of compression we achieve, which translates into making the pipe 10x larger", stated Andy Ben-Dyke, RainStor CTO. "RainStor is a shared everything design where every node accesses all data across the different storage tiers. We use the same method of access regardless of where it is stored, so moving the data between DAS and NAS is straightforward and seamless."