Data Lakes: 8 Enterprise Data Management Requirements
At the EMC World 2016 conference this week, EMC previewed an all-Flash implantation of an EMC Isilon storage system, code named Nitro, that will scale out to provide access to 100 petabytes of storage.
Chirantan Desai, president of the Emerging Technologies Division of EMC, says Nitro will be based on a blade platform configured with 15TB solid-state drives (SSDs). It is scheduled to be in beta later this year and generally available in 2017. Desai says Nitro will sport transfer speeds of 1.5TB per second across over 400 nodes.
All told, Desai says Nitro will both increase the amount of bandwidth available and the total capacity by 50 percent, while reducing the total cost of deploying network attached storage (NAS) systems by 50 percent.
Nitro is designed to advance an EMC Data Lake 2.0 strategy built around the OneFS file system that EMC is positioning as a complement to platforms such as Hadoop. In addition to providing a common global name space, many IT organizations are opting to deploy Hadoop alongside OneFS to simplify deployment and management of a Big Data environment.
As Hadoop becomes the primary data repository in enterprises, EMC is clearly looking for ways to stay relevant in those environments. One way to do that is to provide a storage management platform capable of scaling out without negatively affecting the performance of Big Data applications.
