Albireo Virtual Data Optimizer (VDO), which will be available in the second half of this year, offers inline deduplication and optional data compression as a component of storage systems. Permabit sells a higher-end version of its Albireo deduplication product to NAS vendors such as BluArc, Xiotech and LSI Engenio.
The new VDO application is targeted at smaller NAS vendors that use commodity hardware to create inexpensive and easy-to-use NAS, such as Overland Storage, NetGear, Cisco Systems' Linksys and QNap Systems.
Albireo's index engine works by allowing application servers and NAS systems to quickly identify block- or file-level data already stored on an array of disks. Using the SHA-256 hash function, the software assigns a unique identifier to chunks of data being stored. Those hash identifiers are then stored, along with metadata from the storage systems -- such as the block address -- in an index.
As new data comes in from application servers, it is automatically compared with existing data. If duplicate files or blocks of data are found, the storage system is notified and the data is not stored. When data read requests come into the NAS appliance, the data is delivered in the same way a data snapshot is copied on a backup system, through multiple pointers to the data chunks.
"It means there's no rehydration operation, and we're not in the read path," said Jered Floyd, chief technology officer and co-founder of Permabit. "This functionality is not limited to a handful of terabytes of data. It can deliver this across petabytes of data within the same file system or volume."
Floyd said Albireo is capable of delivering 420MB/sec. of sustained throughput using a single quad-core processor, so as not to cause a bottleneck in the data path.
"We'll never be the bottleneck in the storage network," Floyd said. "We feel if dedupe has a negative impact on a storage system, that's an obstacle to adoption."
Permabit is targeting Linux-based NAS systems with its new VDO product because, after top-tier vendors, it's the fastest-growing market. According to Gartner, the Linux-based NAS appliance market has grown from a 6.5% market share in 2006 to 17.9% in 2009.
Floyd said about 60 vendors offer Linux-based NAS systems, which mainly use the iSCSI transport protocol. The systems typically have from one to 96 drives.
Sign up for CIO Asia eNewsletters.