This session by Tim Stelzner of HP was excellent — Mr. Stelzner clearly knows what he is talking about.
First off — HP is moving storage technology towards something called the “Unified Virtual Pool” — all product platforms, D2D, X9000, P4000/EVA will evolve into this single platform. And right now, the X9000 is here today for file level, and Storage Virtualize Services Platform (SVSP) is available for block level storage requirements. This Unified Virtual Pool will support all protocols — CIFS/NFS, FC/iSCSI/FCoE/SAS/IB.
Next, HP is driving continuous quality improvements with the following processes:
- A permanent ongoing closed loop quality system
- Driving field uptime as a core R&D performance metric
- Daily review of every single support call with data unavailability
- Prioritization of R&D resources based on field down time caused by issue
- Weekly review with senior executive management on the top drivers of down time and the progress in resolving them.
- Ongoing commitment to daily triage, follow-up, investigations, etc.
Why was this news? I don’t know — I would expect that this would be common for a company of HP’s size. Still, it was interesting to see a little transparency. On the other hand, it sure would be nice if HP would publicize their internal list of “lab defects” so that we wouldn’t have to open cases with support just to find out if a problem we’re experiencing is known or not. Also, if this list was publicly available (like EMC’s is) then there could be community discussion and far more support available. Come on, HP!
Then there was this news which is fairly exciting — the EVA would now be clusterable. This would supposedly increase utilization of storage by 300%, and make it possible for server administrators to manage three times more storage and twice as many servers.
An EVA cluster starts with two EVAs (6400 or 8400) with a minimum of 8 drives in each to start. A fully configured EVA cluster scales up to 1.9PBs in 1920 drives in six nodes. And if you need an even larger cluster, you just call HP up who will do some customization for you. Here are some more bullet-points:
- stripe data across 6 controller pairs
- higher availability with network-based replication
- managed through Command View
- operates as a single device
- better software
- thin provisioning
- online migration (for example, for tiering storage)
If one half of the cluster fails the EVA failover occurs without any interruption. And an unlimited license for Volume Manager is included when you buy the cluster.
Now this is very nice, because Volume Manager can then be used to connect in SANs of other manufacturers. So now the cluster also becomes an abstraction layer for other SANs. Volume Manager is licensed for these other, abstracted SANs, in 1TB chunks.
Oh,and apparently, if you’re in the US or Canada, you can ask about the “EVA SW All In Promo” — buy all 4 EVA SW titles when you buy your EVA and get unlimited LTUs for just 30% more than buying just the straight CV unlimited.
Oh boy! Is anyone else as excited as I am about this? I’m running out to the store right now to pick one up. (EVA 4400 Starter Kits are around $20,000 MSRP