Join Newsletter
Trusted Business Advisors, Expert Technology Analysts

Taneja Blog

Taneja Blog / Data Center Systems

Is HP StoreOnce ready to store, once and for all?

HP did quite the news push last week with their latest StoreOnce launch. In the news, the high points revolved mostly around some tremendous performance claims, and a few observations that HP looks poised to take a run at a unique first in the market. That "first in market" claim is a major vendor having the same deduplication algorithm in solutions on both the source side (commonly the backup software or a server agent) and target (commonly a Disk-to-Disk NFS or CIFS appliance or Virtual Tape Library).

As a backstory, HP also made promises that those solutions will some day soon make it possible to move deduplicated data back and forth from source-side repositories and target appliances without restriction. Notably, that level of interchange isn't quite there yet, but the algorithm integration looks to be complete, with significant sophistication around the sticky challenges of index and metadata management, and this looks to make the interchange an easy win in the near future. Moreover, the source side solution is the HP Data Protector client, and implementing the algorithm there seems to have tackled the trickiest area first. Now spinning up other StoreOnce virtual appliances or any other form factor seems like a cake walk.

Why isn't this integration there yet? As pure speculation, I think it is mostly a matter of logistics. Moving that data back and forth between backup software and an appliance requires some sophisticated hand-off and data integrity control. The backup software needs to confirm that all data is in the StoreOnce repository, and it needs to confirm this efficiently. Then the backup software needs to stop thinking about it as deduplicated data, yet likely retain the option to stream it back and then use source-side horsepower to reinflate if the user so desires. And then HP needs to think about how the backup software ecosystem may interact with any of these mechanisms, and test for hiccups. All kinds of interesting possibilities here, but enabling the right workflows is the next step beyond the technology integration, and a complex logistical undertaking.

But just like all the other buzz, I'll give the announcement my "gee whiz, good job" from the propeller head side of things. But I'll also point out there's a bigger story here, and in my opinion it is one that should have the HP competition worried.

You see, the real story, is that this is a major coup for HP in terms of cross-product integration and proprietary IP development. In fact, I'd go so far as to reckon we could at least argue for a long time, if not definitively prove, that this is a bigger and more significant integration than HP has ever won a victory on before.

First, the task of integrating a single algorithm through product domains as big as backup target appliances and backup software is no mean feat. The respective technologies represent plays in markets that are some of the biggest slices of the overall storage industry market. And neither technology is greenfield or immature. StoreOnce has quietly been marching along for 18 or more months now, in the form of multiple different appliances, and HP Data Protector has been around longer than my 5 or 6 years as an analyst with Taneja Group. Moving similar product teams and finding similar technology integration points in similar sized operations has undone the best efforts of many other BIG name technology companies.

But it also isn't just two product integration, it is also cross-product vision integration. HP just turned a long standing one-trick pony into a versatile technology for the scale-out, cloudy infrastructure. Deploy it anywhere, access it from anywhere. Utility-like. Ready to integrate into pay-as-you-go services, conceivably, with more enablements likely to come in the future. The execution seems worthy of the original Bill and Dave, and the B6200 moniker, except maybe back in those days it would have been 6200B.

HP has just shown that they can out play the competition by leveraging unique IP with more precision than anyone else (another interesting tribute to the company's roots, but I'll resist the temptation). And therein is the biggest missed point. HP has just played a hand built on well-engineered extensibility - and the lack of extensibility has long undone the infrastructure, especially storage, and the many vendors that try to innovate around infrastructure. The infrastructure vendor's answer, point products. The StoreOnce answer? Extensibility, integration, vision, and execution.

HP's been suffering for months from the HP-induced turmoil around WebOS. Why? Because the market was excited, and thought a broader integration of HP services - an integrated HP experience across many devices - was about to be unleashed on top of a technology that was unbeholden to any other vendor; maybe just a momentary experience, but it would have been an HP dominate the world moment if you would. Look closer. StoreOnce. Better than WebOS. More important than WebOS. Going to succeed where WebOS didn't stand a long term chance (sorry WebOS guys). But I think it is just a first glimpse of what will come from StoreOnce, and what may come from the variety of technologies that HP is now sitting on. If executive management can stay out of the way.

StoreOnce: HP's claim that innovation is built here. I'm looking for a response from the competition; don't have one yet. If you're an HP customer, take a moment and feel good about the future potential of your infrastructure investments.

Good job Donatelli and StoreOnce team. Power on. And the rest of you guys too - there's lots of peripheral DNA in this one (wink-wink to the vision-influencing 3PAR and P4000 guys especially).

  • Premiered: 12/05/11
  • Author: Taneja Group
Topic(s): HP StoreOnce


There are no comments to display. Scroll down to leave your own!


Leave a Comment

You must be logged in to comment. Click here to log in or register if you don't have an account.