HDS Employs FPGAs to Unify Data Center Storage

Slide Show

Ten Recommendations for Simplified, Intelligence-Based Storage Management

Hitachi Data Systems (HDS) this week turned its block storage systems into a unified storage platform by adding support for modules that make it possible for those systems to also function as network-attached storage (NAS) systems.

By using field programmable gate arrays (FPGA), the storage vendor has been able to create modules that plug into a Hitachi Virtual Storage Platform (VSP) G series based on x86 processors, according to Bob Madaio, director of product marketing for infrastructure at HDS. As a result, HDS is able to take advantage of x86 processor economics in combination with FPGAs to offload other functions onto dedicated processors. This approach enables HDS to shrink the overall physical footprint of storage in the data center, says Madaio.

In addition to adding NAS support, HDS has added tools that make it simpler to migrate data to the cloud, while at the same time tightening integration with VMware vSphere Virtual Volumes software.

Finally, HDS is making available Hitachi Infrastructure Analytics Advisor software that uses analytics to identify performance bottlenecks and storage capacity issues.

Madaio says the combination of FPGAs and a blade storage platform reduces what previously would have taken 12u of space in the data center to five. In addition to saving money on the overall amount of physical space needed in the data center, the amount of energy consumed using this approach is significantly less, says Madaio.

The storage paradox in the data center these days is that while the amount of data being stored is increasing, the amount of physical space required to hold it all is shrinking. IT organizations need to save money on both space and energy to help offset the cost of additional storage. The challenge, of course, is going to be figuring out how to move all the data that currently resides on a separate NAS system into a platform that finally unifies the management of all that data.