Days are numbered
Is this the first nail in the EVA coffin, think its days are numbered, are HP even going to release another generation of EVA?
HP killed EVA clustering sales at the end of 2010, meaning that scale-out virtualised storage capacity across EVA arrays and thin provisioning are no longer available. 3PAR products fill the gap. LSI Corporation is closing its Storage Virtualisation Manager (SVM) product line down. HP EVA clustering, launched a little over six …
"Is this the first nail in the EVA coffin...." Sorry, you're trying to jump the queue there! Competitors have been insisting on nails in EVA's coffin for years, yet so far none have managed to get one to hammer home. Having said that, the 3PAR buy does seem to kill any chances Wade Dolphin had of pushing it up the foodchain, and the hp MSA/Lefthand combo seems to be eating it from below....
".....are HP even going to release another generation of EVA?" According to the roadmap I've seen, the answer is "yes". The EVA product is much more than the clustering bundle and still seems to be selling reasonably well.
Fun story - when the hp salesgrunts came to do a roadmap talk to us about the EVA a few years ago, we pointed out that the proposed EVA cluster overlapped with the external storage capability of the XP array, in particular, the SVS200, which was kind of like a diskless XP array. The scarey part was the hp EVA team weren't aware of the SVS200, a product from another hp storage team!
SVSP was sold prior to EVA cluster as a standalone HP product, they just made it an off the shelf purchase by adding in the EVA's in underneath, it also allowed the EVA's to scale beyond their current hardware limits without the need for a product refresh. LSI discontinued support for SVM a while ago, only the HP exclusive OEM kept SVSP development alive. So I think it's safe to assume HP ended the OEM relationship due to the 3PAR purchase. With the end of the OEM SVSP and the EVA cluster were canned, but EVA remains as a product in it's own right, with at the very least another generation on the cards.
As I see it then HP currently has no choice except to deliver a new EVA-generation - the hardware refresh is actually in-line with the historic lifecycle of approx 3 years. Features (thin provisioning, online lun migration etc) that got announced at end 2009 and later pulled (but were shown later at eg. Tech Forum in Vegas in 2010) got delayed by lack of R&D because of the crisis I guess.
The point is that HP has a very massive customer base on the EVA which is happy and very satisfied with it. HP technicians and partners are abundant for support - none of that exists yet on 3Pars. Those customers will be expecting at least one more generation (due in march) while HP learns the new product. If HP was to simply drop the EVA so its no longer available then whats to stop those customers from going somewhere else? -It would (maybe) create a too competitive situation for them making it difficult for HP to capitalize on the 3Par investment as EVA customers generally would go for a cheaper product in the current market - 3Par is only really interesting if you buy a lot of the options and then it is definately not a cheap product, despite intelligent tiering to large capacity but slow disks.
The major factor in both the EVA and 3Par product lines is that the backend is all based on fibre channel - and the disk vendors are mostly developing on -and producing SAS-based drives, so currently those systems that can do block-level virtualization, fibre and iSCSI connectivity on the front-end and also scale out while using SAS as a back-end are winning the market - eg HDS and IBM. So thats what everybody is waiting for - thin provisioning, SAS-backend for the EVA to make the disks cheaper and the possibility to move data transparently between disk groups, if HP delivers all of that in march then a lot of customers will be happy to buy and stay on the new EVA generation since a lot of them really dont feel the need for a lot of the advanced features from 3Par which was developed for hosting purposes and not for the broad market.
After that then what both the EVA and 3Par product really needs is to be recompiled on an x86 CPU so controller revamps become easier and the controller boards cheaper.
The greatest mistake that HP has done in all of this that they have not been honest and fair towards the existing and large customer base for EVA nor their partners - HP could have squashed all this doubt from day one by coming clean.
Two things, I think. One is that any future product line based around fibrechannel networking is clearly regarded as no future at all. Since the EVA is HP primary go-to market for legacy storage markets LSI felt it had no future with the software.
Second, HP much touted guarantees to continue with the EVA must look suspect now. It's hard to believe that the EVA has any future beyond it's current feature set and that 3Par is the long term future. Trying to make the point that supporting the existing EVA is a requirement is as useful as spitting into the wind.
".....any future product line based around fibrechannel networking is clearly regarded as no future at all....." Why? After all, the last thing I want to do is put my storage traffic onto the same cable as all my LAN traffic like the iSCSI and FCIP crowd suggest. I don't want my core apps waiting on data because some cretin has decided to dice up the VLANs and stolen the bandwidth I need, I want my SAN traffic going over dedicated cable. I'll keep iSCSI for the edge apps (and even then I can put an iSCSI header on an EVA, XP, or just use an hp Lefthand array). That dedicated cable means FC for now, especially as it seems we can carry on expecting a few speed bumps for a while. The interesting bit will be if SAS networking plays out and whether it can replace the majority of the existing FC SAN cabling and switches, but I think that would mean a lot more work on the SAS protocols (any SAS gurus know better?).
Just to remind you, some VP at NetApp said back in 2007 that SAS would have replaced all FC by 2010 - it seems he was a little premature!
What I don't really want is a fibre back end on my array as that does seem to be the expensive choice what with SAS disks going to get faster and more power-efficient than FC disks, and being compatible with cheaper SATA too. As hp have already shifted the MSA lines to SAS it would seem logical to suppose the next gen EVA will be SAS too, especially as the EVAs always seem to use MSA disk shelves.....