A shot across the bow
HP’s acquisition of Polyserve is a ~$250 million (my SWAG, we’ll have to see what, if anything, gets reported on the 10K) bet on the future of storage. And I think it is a good one.
HP needed to do something. Their external storage systems business has been flat while EMC has been taking share. They still have a strong number two position, but that is not the way to win Mark Hurd’s love.
Polyserver: commercial storage cluster
I haven’t devoted much time to Polyserve, but at a high level I like what they are doing. Polyserve takes the NAS storage cluster concept right into the heart of commercial applications. They support Oracle databases (learn more at Kevin Closson’s excellent blog), SQL Server and DB2. This is where the big iron storage boxes live, as well as their less-costly mid-range cousins.
A good overview of the buy and HP’s reasons is at Red Herring. Yet the importance of the acquisition goes well beyond HP.
Cluster storage land-grab
The real impact is on every other major storage player. As I’ve noted, storage clusters aren’t coming, they’re here. HP’s move just underscores that fact.
The wide-awake storage players are now putting out feelers to buy or partner with every decent storage cluster technology play. If you have a storage cluster startup and don’t get a phone call from someone in the next month, maybe your stuff isn’t all that cool. Or your genius isn’t appreciated by a clueless world.
EMC already has investments in a number of next-gen startups – such as my former company YottaYotta – which they’ll need to take to the next level by either licensing or acquiring. Since EMC’s storage business has been growing at a healthy clip, they may not feel the need to act fast, leaving it to competitors to claim the high ground.
IBM always has at least a dozen options on a low-boil, but deciding to pull the trigger on one and the plug on another just isn’t their thing. With their blade server leadership it would seem a natural direction, but they punted on storage arrays too.
Sun is well-positioned to use x4500’s with Google-style clustering to create low-cost, high data integrity storage clusters. If the Solaris group takes it on, it could be good. If they don’t, well, don’t hold your breath.
The StorageMojo take
Bravo to HP for buying Polyserve. If they roll out the products and support in a timely manner they will steal a march on everyone else in the industry. Their biggest problem will be selling Polyserve to their sales force, a skill HP has yet to master. Lower ASPs always give sales people the willies so it is vital that they feel like they are taken care of. Like Sun’s mis-marketed x4500 the Polyserve products could find themselves in a completely undeserved oblivion if HP’s storage group just tosses them over the wall to sales.
Comments welcome, as always. I’m traveling until Thursday so I may not moderate as fast as I normally do, but moderate I shall!
“The wide-awake storage players are now putting out feelers to buy or partner with every decent storage cluster technology play. If you have a storage cluster startup and don’t get a phone call from someone in the next month, maybe your stuff isn’t all that cool. Or your genius isn’t appreciated by a clueless world.”
or you invent/operate outside the universe that is the US of A…. perhaps the big boys cannot afford the transatlantic call costs!
Jelloknee,
Hey, if you are outside Silicon Valley, Route 128 or metro Denver you face some extra hurdles. There are a few storage companies outside those two areas: Isilon, Texas Memory Systems, Data Direct, but not many.
It is kind of like wanting to be an opera star while staying in Tulsa or – for my UK fans – Glasgow. It is too bad, but if you think you’ve got the opera mojo you need to go to New York, Milan, Paris to strut your stuff.
Come to Silicon Valley, Jelloknee! The weather is better, the VCs more plentiful, and the resources available to startups are unparalleled. If you make it you can make it huge, and if you don’t there is no shame in startup failure.
Robin
Thanks Robin,
Disheartening advice but the reality I suppose.
I do find the term ‘startup’ as used in the valley highly amusing. Such ‘startups’ receive millions to build software that will be over-priced and under-spec’d just to pay off their investors. The unparalleled resources can simply end up being a distraction and not a useful tool.
Surely the way to true satisfaction is the hard way..
– Build a product on a shoe string with prolonged periods of no string.
– Innovation through starvation can be both very effective and rewarding (not in the financial sense though!)
– Build an ethos that is based around re-use not re-invention.
– Build a product that is tied to an OS and not proprietary tin.
– Build tough skin … and lots of it. Prepare to encounter and counter the phrase:
“Who are you, never heard of you, goodbye”
– If your lucky enough to get a customer or two to believe in your vision (and you will need buckets of luck) hold on to that customer, love that customer, nurture that customer and remember, that customer is way more important to you post-revenue then pre-revenue.
– Plan to build on success…. Just because you have a couple of great customers its not time to paaaarty, two customers will just about pay for the fancy Nespresso machine.
Whether in smoggy California or sunny Wales, if you manage to survive the above with friends and family still talking to you and the business sustaining itself through further customer success then satisfaction and possibly a trade sale will surely come.
If not then your screwed 😉
Too interesting a topic not to comment on.
With Isilon’s slams against NetApp, NetApp’s (curious) response’s, HP’s move to acquire Polyserve, and subsequently Cisco’s move to acquire NeoPath, there is certainly a lot of interest and movement in this segment of the storage industry lately.
As for storage clusters in general, HP is looking more and more desperate in this area. Remember Compaq’s ENSA? That was a really good vision, with some decent amount of reality behind it for grid (like) storage. Compaq even had the people to execute on that strategy. My how things have changed, HP decided to execute the strategy by killing it. Now 5 years later they decide to spend a few 100 Million $’s to re-acquire technology they could have had 3 years ago.
IMHO, Polyserve has a looonnnnggg way to go before it can become mainstream NAS. It has some of the scalability capabilities of NetApp ONTAP GX, Isilon, and others (though not all), while having all the shortcomings. They don’t have snapshots (wasn’t that a requirement of NAS since about 1997?), don’t have replication, don’t have integrated NFS and CIFS, etc. etc. In fact, the lack of Snapshots and having to choose NFS or CIFS really puts them back into the state of the art of NAS circa 1995.
Even fourth place NAS vendor Sun has integrated snapshots with decent NFS and CIFS.
For niche, high-end HPC environments, Polyserve NAS will be great, though best I can tell, that isn’t a large part of HP’s market.
So, looks like HP will get some good technology, and a lot of possibilities, but will have a lot of work to do in order to match the big two in NAS. To date, investing in storage engineering development (vs. acquisitions) is not something many of the larger systems companies have done much of, so I think the jury is out for quite a while…
–aSG
To anotherStorageGuy:
There are several inaccuracies in your arguments – for example, both CIFS and NFS are available on the Linux version (NFS is scale-out, while CIFS is Active-Active but not parallel) and you are similarly half-wrong / wrong regarding other “missing” features.
Disclosure: Like anotherStorageGuy, I’m not impartial either.
I agree Polyserve has a long way to go before they can compete with Isilon in the NAS market.
Polyserve does support CIFS and NFS on their Linux product.
– This product DOES NOT have full windows ACL support, which is what most customers want for CIFS.
– By Polyserve own admission, they are focusing on the Windows product, because most of the growth is in that market.
– As noted earlier, it is not scalable for both protocols.
Bottom line is that they offer both protocols at a very basic level.
Polyserve does have snapshots at some level.
– You can use the snapshot capabilities of the HP EVA. Polyserve integrates this functionality in their interface to act like it is a Polyserve snapshot
Or
– You can use Microsoft VSS for snapshots if you have the windows product.
Bottom line is that they do not offer POLYSERVE snapshots, they only dress up someone else’s snapshot capabilities.
Polyserve supports replication
– You must buy a third party product.
– That 3rd party product supports only host based replication.
– It is not integrated into the product.
For a NAS solution, Polyserve is years from having the feature set of Isilon. I’ve used both products extensively, and HP will have to make some serious investments to make Polyserve a real player in the High End NAS Market.