Permabit branched out from its archiving roots today with the introduction of data deduplication software designed specifically for capacity optimization on primary storage devices. The software-only Albireo solution is targeted at storage hardware and software OEMs. Permabit claims to have a number of design wins, although OEM products based on Albireo software aren’t expected to appear until late this year.
Permabit hopes to differentiate Albireo from existing techniques for optimizing primary storage on a number of fronts, including implementation flexibility. For example, unlike data deduplication or compression solutions that are designed specifically for file-based data, Albireo can be implemented with block, file or unified (block + file) storage.
In addition, unlike capacity optimization techniques that are limited to either in-line or post-process deduplication tecniques, Permabit’s Albireo can be implemented in-line, post-process or in a hybrid technique that Permabit refers to as parallel deduplication.
Jered Floyd, Permabit co-founder and CTO, claims that Albireo has no performance impact (unlike some of the early entrants in the file-based deduplication market) due in part to a memory-based index that the company claims reduces disk access by 99.9%.
In addition, unlike some appliance-based approaches to capacity optimization, Permabit’s embedded approach does not introduce a single point of failure, according to Floyd, thus guaranteeing data integrity.
Although Albireo is a deduplication-only solution, it could potentially be used in conjunction with data compression-based solutions.
Albireo is based in part on Permabit’s existing Scalable Data Reduction deduplication technology, which enables scalability into the petabyte range, according to Tom Cook, Permabit’s CEO.
Deduplication ratios depend on a variety of factors, but with fixed-chunk user directories (e.g., Microsoft Office documents) Permabit claims deduplication ratios of 2.8:1 with 4KB blocks and 2.7:1 with 64KB blocks. With variable-chunk user directories, Permabit cites deduplication ratios of 3.9:1 (4KB blocks) and 3.8:1 (64KB blocks). And in VMware environments, the company claims typical deduplication ratios of 36:1 (4KB blocks) and 26:1 (64KB blocks).
In terms of performance, Permabit claims more than 140MBps per CPU core with 4KB blocks.
Other vendors with solutions for primary storage optimization include EMC, Exar (via its acquisition of Hifn), greenBytes, NetApp, Nexenta, Ocarina Networks and Storwize.
Permabit delivers Albireo to OEMs in a software development kit (SDK) that OEMs can use to integrate the data deduplication software with their existing stacks.
