French company Seanodes has introduced software architecture that it says allows companies to leverage unused processing power in its storage infrastructure to create a virtual storage pool.
The introduction of this application, the firm claims, has in effect created a new category in the storage space that it is calling “Shared Internal Storage.”
Companies have always had a certain amount of internal storage capacity in conventional SAN (storage area network) and NAS (network-attached storage) network storage architectures that go unused or are underused, Frank Gana, director of business development told LinuxInsider.
“The cost of not using this capacity can add up significantly for some companies,” he said.
Aggregating the Space
Its product, Exanodes, is based on a concept similar to VMware Xen or Microsoft’s Virtual Server, which aggregate and consolidate CPUs in application servers. Exanodes, by extension, aggregates and consolidates storage devices in application servers. It does this by enabling a system’s internal disks to be managed as a single shared storage pool. The application runs on Linux, and a version for Windows Server is in development, said spokesperson Dan Miller.
Exanodes also supports ultra-fast RAID (redundant array of independent drives) rebuilds, according to the company, by intelligently distributing the RAID set across multiple network nodes so that a terabyte of data can be rebuilt in under an hour, minimizing the risks of a second disk failure during the rebuild window. A similar rebuild with conventional disk arrays can take days, it said.
Thus, the need for dedicated storage servers is reduced as each node can function as both an application and storage server. The application also can help the system maintain a sustained read or write performance of several gigabytes per second.
Exanodes works with all block storage devices, hard disk drives, SSD (solid state drive), RAID and external DAS (direct-attached storage) on the market today, according to the firm.
For the end user, Exanodes will mean reduced management costs and less need for additional external network storage arrays. Other benefits include smaller data center floor space and reduced power consumption.
From a system perspective, the end result is an aggregate formatted capacity made available to all the application computers, according to Laura Dubois, program director of storage software at IDC. “No stranded capacity exists, and there is no single point of failure, since mirroring across computer systems would be employed.”
Beyond these generalizations, though, it can be difficult to quantify the bottom-line savings a company would realize with such a deployment. “Much depends on the needs of the company, its size and existing infrastructure,” Gana said.