In-Depth
The Coming Storm
Does the current recession bring inventive and innovative storage strategies with it?
Local newscasters are all a-twitter about a burgeoning hurricane in the Caribbean that has a remote possibility of making landfall near Tampa Bay, FL. While the thought of a Katrina-like storm sends chills down the spines of most Floridians, this one doesn't seem ready to roar. Still, backups are being made more frequently as the storm season gets underway and plans are being reviewed to move to alternative sites should power or communications outages occur. The "gathering storm" is an overused metaphor, of course … it is, however, appropriate given current trends in storage.
According to recent user polls in several trade publications, NetApp and EMC appear to lead the pack in several categories: sales competence, initial product quality, product features, product reliability, technical support, and the oft-asked "would you buy this product again?" category. Without a drill-down into the demographics of the respondents, there is no way to be sure what these results really mean. It is probably safe to assume that these polls reflect the opinions of large enterprise users rather than IT professionals in small and medium enterprises.
In fact, of the brands tested, most were venerable big iron vendors. 3PAR was the only "newbie" included in the questionnaire.
If only considering large shops, the data seems to confirm what we have reported here. In the majority of large enterprise shops, storage remains a peripheral matter. It is largely viewed as a repository for undifferentiated and anonymous data by the folks in IT. Capacity management is the order of the day and any box of spinning rust will do just fine, provided that it is backed by a reliable maintenance contract with one-throat-to-choke guarantees.
In many of the Fortune 500 companies that I visit, storage acquisition deals are being done in the front office -- by the CFO or purchasing department, and not in the back office where IT lives. A representative from a large financial institution recently told me that he would like to purpose-build his storage and to buy what his applications really need, but senior management has already struck bargains with a handful of vendors and his selections for products are limited to what's on the approved list.
When IT is involved in decision-making, the brand-name hardware vendors are cutting prices (software maintenance costs kick in next year) and proffering a twist of their own on the "do more with less" value proposition. They begin by sympathizing with IT managers on the cutbacks in their CAPEX and OPEX budgets, noting the staff reductions that have occurred -- or are about to take place -- given current economic conditions will make it difficult to deliver the expected improvement in service levels that the front office expects. The vendor then says that their box, with its "value-add" software components, will save the day -- enabling the consumer to do more with less.
Value-add is a blanket term that increasingly includes stovepipe management software, stovepipe de-duplication and compression software, stovepipe thin provisioning software, stovepipe array virtualization software, proprietary data path-ing and RAID schemes, and stovepipe continuous data protection software. These features are very valuable -- to the vendor, at least. Basically, they help to lock the customer into Brand X products and make it more difficult for Brand Y vendor to dislodge them. That is what stovepipes are all about.
Such a value proposition is extraordinarily compelling. Buying everything in one box has the appeal of pre-integration. Vendors argue that, since IT doesn't have enough bodies or bandwidth to cobble together its own solutions from best of breed components, the pre-integrated one-stop shop is the next best thing. Pre-integration makes interoperability issues the vendor's problem, they say, and not the IT department's.
This is not true. All the automated management and "phone home" capabilities in the world cannot be relied on to keep a storage platform up and running and data safely stored out of harm's way. The more complex a platform becomes, the more vulnerable it becomes.
It's only necessary to recall what recently happened when VMware joined the growing list of software vendors who blew their pre-integrated value proposition to smithereens when they issued a patch that shut down most ESX 3.5 servers. The same thing has happened (and will happen) to those value-add software stacks vendors are stacking up on storage platforms today. This has to raise the question of whether there is any real value to vendor pre-integration or is it just a way to charge a lot more for a "signed" hard disk.
When a failure does occur, it is almost a certainty that the consequences will be shouldered by IT -- and not by vendors. The longer the vendor takes to resolve the problem, the more the IT professionals will be criticized and their reputations damaged.
The alternative is to design your own solution. That means: buy affordable and resilient bare-bones storage components that conform to some sort of "unified management paradigm" -- whether this is an SRM strategy based on software from Tek-Tools, CA, or a handful of other vendors, or the latest idea from the W3C: Web services-based management.
With the W3C approach, value-add service components, previously delivered on storage hardware stovepipes, are added to infrastructure at a network layer, using platforms from F5, Crossroads Systems, Tervela, or any number of other hosting appliances. As Web-based resources, they can be leveraged by the data from applications that are already "Web-services enabled" (little software houses like Microsoft, Oracle, Sun, VMware, etc. already have Web services functionality built in).
The result is an "economy-of-scale" usage model for the storage services (de-dupe, compression, encryption, etc.) you really need, and without the cost of software maintenance support contracts on "value add" software that you don't use -- but that the big iron vendors sell you.
This sounds like a lot of work. Do it yourself (DIY) usually is. That is what the vendors are arguing these days at conferences and seminars. In their words, "If you are short handed already, where are the resources going to come from to help you cobble together the DIY solution you want to build?"
They make a good point, and one I might have grudgingly conceded a few years ago. Truth is that a lot of storage integrators went the way of the dinosaur back in 2001, mainly because the dotcom bubble-burst saw many of the three letter vendors taking key accounts away from their channels and moving them into direct sales. The result was that integrators had to release their qualified integration staff because there was no more money coming in to pay their salaries. Many became little more than aggrandized order takers.
I am, however, seeing evidence that the landscape is changing for the better. Recently, I got to rub elbows with the smart people from reseller/integrator Nth Generation Computing in Anaheim, CA. A few weeks before that, I was a guest of VION, a well established Federal integrator now making a foray into the commercial space. Both companies seemed to have a great ofhandle on storage issues and considerable competence in build-to-fit storage solutioneering.
In addition, I have been fielding inquiries from vendors asking me to make introductions to qualified storage integrators and engineering firms, such as the high availability storage engineering firm, Zerowait. The vendors are seeing the wisdom of having qualified integrators create pre-integrated solutions out of best of breed, hardware-agnostic, storage software building blocks.
I am also seeing vendors of hardware building block components, like Xiotech, with its Intelligent Storage Element (ISE), spending considerable effort to grow ecosystems with best of breed software providers. Such relationships lead inevitably to plugfests and reference models that help reassure consumers that they don't need to settle for overpriced wares from three-letter firms to get the functionality -- or the support for the integrated solution -- that they require.
For IT planners, this means that they are not alone in the DIY game. A storage integrator can become their trusted ally once again. The integrator can take some of the load to help build and support tailored solutions at a fraction of the cost of buying one-size-fits-most products and with the same or better support model. This, in turn, has the effect of reducing the overall cost of storage acquisitions and building manageable solutions that are truly customized to business needs.
In truth, the current recession might be pushing some companies into the false security of one-stop-shopping. From where I'm sitting, however, necessity is also beginning to foster inventive and innovative storage strategies that will save the company money, reduce risk, and contribute meaningfully to productivity and top-line growth.
Even some of the brand vendors are starting to worry about this gathering storm of economic constraint and innovative thinking. Last week, a regional manager from EMC was falling over himself in a conversation with me to channel his company's view of the future: services to help companies field the right infrastructure to support business processes – even if EMC's brand isn't on every box.
I am still recovering from the shock, but you can send me your views at [email protected].