NetApp duplicates deduplication
Two much of a good thing
Comment By buying Engenio, NetApp gets engineers working with Permabit's Albireo deduplication technology to increase the effective capacity of Engenio's arrays. Trouble is it looks better than NetApp's in-house A-SIS dedupe. Oops. What does NetApp do now?
A-SIS works with Data ONTAP, the software driving NetApp's FAS arrays, and is near-primary deduplication. Its use is not recommended for highly-active data due to data I/O slowing down.
Albireo is a dedupe technology developed by Permabit that is claimed not to have any performance impact and not to require data rehydration separate from the normal process of reading files – or data – from hard drives in arrays. It is being used by BlueArc and Xiotech who aim to add deduping functionality to their storage products without losing a performance edge. Ditto Engenio.
Now NetApp has two dedupe technologies and they would appear to be incompatible. What are the options?
One is to dump Albireo. That would appear to remove a future benefit of buying Engenio arrays: the ability to hold more data than the raw capacity would suggest, thus making them better value.
Another is to replace A-SIS with Albireo. This would provide a consistent deduplication environment and enable greater integration of the FAS and Engenio storage enclosures. It might take a while as the ONTAP storage stack is a multi-layered beast and integrating third-party technologies into it has taken time in the past.
The third option is to stay with both and avoid making an either/or decision. This would limit NetApp's ability to sell FAS arrays into Engenio shops as there would be two pools of deduped data, but it is probably workable.
And then there were three: a complicating factor is that NetApp is now selling a Fujitsu Virtual Tape Library in Europe which uses Quantum's DXi deduplication. We can say two's company but three is a crowd.
This provides a fourth option: dump both A-SIS and Albireo and go with Quantum. However DXi is not positioned as primary data deduplication technology which rules it out.
There is a rumour that EMC is developing its own primary deduplication capability and something could be revealed at EMC World in Las Vegas next week. With EMC now being much stronger competition for NetApp, with its VNX range, it could be that NetApp will feel itself forced to pick up the Albireo ball and run with that for FAS so as to have a ready-to-implement answer for whatever EMC does.
HP is developing its StoreOnce deduplication and Dell is going bananas spreading its Ocarina data compression and optimisation technology across its storage product range. These are two more market factors that would encourage NetApp to make its dedupe offering stronger.
An Albireo choice would be great news for Permabit: having NetApp as a direct OEM for its technology. It would also get NetApp out of a looming inconsistency trap and, all-in-all, looks to be the likeliest choice. ®
no global de-dupe...
ASIS (which ain't ASIS any more... it's just NetApp De-Dupe) does not support global de-dupe, i.e. it does not de-dupe across arrays. So, I don't see any issue around the 'two pools of de-dupe' area between the legacy NTAP pools and the new Engenio pools. It is also suitable for all but the highest workloads and is well proven - it is one of the gems of WAFL. I expect introducing another de-dupe product into WAFL would be very disruptive and time consuming. Net, I expect them to leave de-dupe alone, perhaps just pinching some ideas from Albireo. A full replacement? I doubt it.
My bet is that they'll allow both to run in parallel for the time being. No big deal.
It's all about the File System
Chris, your statement about DXi is absolutely correct, this is an appliance designed for Target Based Inline Deduplication. However perhaps you are missing the most important fact.
The file system that sits below every DXi is StorNext. A file system that is absolutely designed for Primary Storage, in fact you could argue that only with StorNext are some high performance environments able to function.
Now I wonder who signed an agreement with Quantum a few weeks ago to sell Stornext?...would that be NetApp?