‘Elastic Storage’ aims at some of the same goals as EMC’s ViPR 2.0 software, which launched last week. The general idea is to be able to access any type of data across all of an enterprise’s storage systems in multiple locations. Just as it did with servers, virtualisation treats storage arrays as a collective resource that can be split into virtual systems to suit application requirements.
A key benefit of Elastic Storage is that it can automatically move less-used data across the whole pool of capacity into less expensive commodity storage, leaving faster and more expensive resources such as flash for more pressing data. That automatic reshuffling, guided by policies and real-time analytics, can cut storage costs by as much as 90 percent, IBM claims.
The technology also eases administration, speeds access to data, and lets enterprises expand their capacity quickly, the company says.
Like ViPR, Elastic Storage is designed to work with storage gear from third-party vendors. IBM says it will work with any company’s products. It supports the OpenStack protocol, allowing users to access and manage data across private and public clouds, and can work with Hadoop and other open APIs (application programming interfaces), IBM said.
Elastic Storage was able to load 200 million pages (5TB) of data into Watson’s memory within minutes, IBM said. The architecture of Elastic Storage allows it to work with thousands of yottabytes, each of which is equal to 1 billion petabytes.
In addition to being offered as software by itself, Elastic Storage will be available later this year as a cloud service through IBM’s SoftLayer IaaS (infrastructure as a service) offering.