2 Replies Latest reply: Jul 1, 2010 10:31 AM by bruce.breidall RSS

Deduplication of backup data using the NetApp VTL

keith
Currently Being Moderated

In the January issue of Tech OnTap, I wrote an article on the unique capabilities of our VTL deduplication. Unlike NetApp Deduplication for FAS technology, that is specifically optimised to deduplicate data in the context of primary storage, NetApp Deduplication for VTL is designed to work with stream-oriented backup data. The NetApp VTL deduplication algorithm is alignment independent so that it can identify data duplication occurring at any offset within backup streams, and can eliminate that duplicate data based on the extent size of the duplicate data itself, rather than being based on fixed size, fixed length blocks. We've also designed the algorithm to be rate adaptive, such that portions of it can switch between inline processing and post processing, depending on workload, to ensure that backups never get slowed down by deduplication.

You can check out the article at http://www.netapp.com/us/communities/tech-ontap/vtl-dedupe.html

What do you think about the NetApp VTL? Do you have plans to leverage this new deduplication technology with your existing VTL systems?

  • Re: Deduplication of backup data using the NetApp VTL
    bruce.breidall
    Currently Being Moderated

    Do you stil stand by what you said, knowing this product is dead? I am seeing a lot of issues with getting the post process done, and a large amount of free space is needed in order for the work to get done, much more than was ever revealed before buying the product. I feel your statement implying in-line dedupe is wrong, because we really have no window, backups run all the time. It can't be avoided in a real shop, depending on your SLAs (like DB archive logs), and I am seeing a definite slowdown when backups are running during the window. These are not big jobs, just lots of small ones. I also see a 100% difference in dedupe savings from one VTL to another, for the same SQL data. The only difference is how it is backed up. One via the host SQL backup agent, and the other with the host native agent. Thanks.

More Like This

  • Retrieving data ...