NetApp has added block-level de-duplication to its Virtual Tape Library (VTL) product line and claims an up to 20:1 de-duplication ratio. At last, the company has an answer to Data Domain, Quantum, and EMC.
Until now, NetApp VTLs used hardware compression only and achieved a 50 to 60 per cent reduction in data volume. With VTL de-duplication code up to 95 percent of data can be eliminated depending upon the redundancy in the backup data.
The de-duplication is not NetApp's ASIS technology as used in ONTAP. It has been designed and built from the ground up and uses variable length strings.
The de-dupe can be selectively applied. Not all virtual libraries in the VTL need be de-duped.
De-duplication is carried out after data is ingested, although incoming data is fingerprinted using a so-called rolling hash to speed later de-duplication. Within twelve months, inline de-duplication will be added together with the ability to auto-switch between inline and post-process de-dupe depending upon the de-dupe rate, so as to meet backup window limitations. This is roughly similar to Quantum's DXi technology. NetApp would like us to think that it's switching between inline and post-process modes will be better.
Another future addition is replication of de-duplicated data between NetApp VTL systems. Today, it is limited to raw data. A third roadmap item is clustering of the VTLs - something we can definitely expect in the future. There are, however, no plans to converge the VTL operating system with ONTAP. NetApp sees performance and scalability advantages through having a separate VTL appliance with its own O/S.
Despite this separation there is some investment protection because the VTL can be converted to a FAS3000 array whose storage hardware it is based upon.
NetApp says it is pleased to be able to offer de-duplication across all its product lines, taking a sideways swipe at EMC which does not. Existing NetApp VTL customers with maintenance contracts will receive the de-duplication capability at no charge. ®