Well they can tell the channel what the hell is going on...
The silence from HP is deafening.....
According to my contacts, as recently as last month (Dec 2010) HP planned to announce an EVA (Enterprise Virtual Array) refresh in mid-Feb. But at the very end of the year, HP surprisingly withdrew EVA clustering and allied thin provisioning. HP's 3PAR is supposed to fill the gaps. Add a refresh delay into the mix and …
The silence from HP is deafening.....
when they forgot to make any continuing investments in storage in the mid 2000s.
Then Dell-the-server-company bought EqualLogic, and HP-the-server-company woke up with a start & bought LeftHand (now the P4000), lurched sideways and bought Ibrix (X9000), and thought that they'd fixed the storage problem that was the EVA.
Wrong choice. The P4000 is small beer, a sort of one-shot technology that was OK when they bought it, but it's not going anywhere, doesn't scale well and isn't that popular as it's iSCSI only (the EVA is FC). The X9000 NAS boxen were never going to set the world alight; Ibrix and their dismal sales record proved that already. That's what you get for buying companies on the cheap.
So HP dig deep into their trouser pockets and buy 3Par, and discover the fit isn't that splendid; the EVA's natural market was at the channel low-end where bundles of storage, servers and networking make sense. 3Par doesn't fit in that sales model; it's a different beast than the EVA.
Data growth & management of the increasing amounts people want to store has illuminated the EVA in the worst possible way; it sucks as a platform for tomorrow, and it's barely adequate for today. 3Par might be the route forward for today, but it's asking a lot of EVA customers to do a rip-and-replace to make the switch, especially as they've already invested and need a return on that money. (Interestingly, 3Par is equally crippled for the long term; no dedupe, no compression, not unified storage, and unlikely to get these essentials any time soon. Thin provisioning is the big claim to fame, along with performance in a straight line. That's it.)
No HP storage visualisation appliance to ease that pain either, since the SVSP has been unceremoniously dumped.
Rock and hard place. Channel partners must be wondering what HP will do next, since they're facing some seriously stiff competition from particularly NetApp, with whom they're equal 2nd according to IDC.
Disclaimer; NetApp employee, and delighted to be so.
I work for a partner of HP's amongst other vendors (NetApp, EMC, HDS included) and
I agree HP's communication around whats happening with EVA has been quite appalling and even when you ask them directly, they won't tell you anything.
3Par is also not the answer to everything, but I am excited by its capabilities, and can see very clearly where it will go due to its architecture
That said 3Par may not have de-dupe or compression but they don't need it as the Thin provisioning capability they do have far exceeds other offerings. Customers typically see 60% - 70% + savings in space (certain enterprise customers are quoted as only having to buy 1TB for every 3TB they used to).
By the same token how much space saving does De-dupe actually give 25-40% I'd guess in a typical scenario (mixed environment, admittedly could be higher in area's like VDI) which is still a valuable saving.
But lets look at it further if I have 50TB today, but after de-dupe will only need 20TB, how much NetApp storage do I need to buy upfront (to use Snapshots as well)?
The unified Storage argument, HP has a range of NAS gateways that can integrate into the solution depending on the workload, the one disadvantage is its not centrally managed which I guess is your point.
As for Lefthand/P4000, yes its a one trick pony but its also one thats eating NetApps lunch from the low end, we have seen a lot of success particularly integrating with HP's Blade and Virtual Connect technology.
Interestingly some of my friends who work at software companies like VMware and Symantec think NetApp will not grow as strongly this year, largely because their technology hasn't moved on that much (though of course I'm not in the know as to whats around the corner), and also because of poor block performance and utilisation and in terms of Cost particularly software, i.e. the value that certain software titles give versus the benefits isn't great especially for dual head configs.
In terms of Storage virtualisation, you forget that HP also have the P9500 from Hitachi which can perform block based virtualisation, NetApp V - well thats just a volume manager really, you could do the same with a Solaris box and ZFS, or a Windows Storage Server....
From a HP perspective I'm probably more worried about what Dell will do with Compellent (if they can execute as well as hey have with Equalogic then they will be a threat), Of course EMC, NetApp and HDS will continue to be a threat until HP do finally tell us all whats happening!
The reality is that channel partners don't worry too much so long as they have something to sell i.e. a solution and they can make a decent margin on it, now end customers is a different kettle of fish
Biting the hand that feeds IT © 1998–2017