Integrated IT stacks are bundles of storage, compute hardware, networking and software being sold by groups of manufacturer “partners.” The value proposition should sound pretty familiar,
EMC, Cisco and VMware are partnering to provide the Virtual Computing Environment (VCE) Vblock, a collection of storage, servers (from Cisco), switches and virtualization software designed to drop into a data center, ready to go. NetApp has partnered with VMware and Cisco as well to create the FlexPod, a similar offering. Oracle is doing its own stack with its Sun servers and storage, plus Oracle database software and its own virtualization solution. HP, too, is offering a bundle with its new 3PAR storage, ProCurve and HP servers and software.
This concept certainly isn’t new. Bundling is a primary value proposition that storage integrators have been using from the beginning. Getting all the components from a single source and having that source take “systems responsibility” has near-universal appeal in IT, as well as other industries that rely on complex systems. As another example of a systems solution, integrated IT stacks certainly have value.
Like the “appliance” format, they simplify implementation and eliminate the potential problem of open systems components not playing nicely together. The appliance format has become somewhat of a standard practice for releasing new technologies, as it allows companies to sell their new products more easily, without an army of professional services people. And implementation is becoming a bigger and bigger issue with IT organizations, which have fewer experienced staff members to get new products up and running. Support is the other aspect, as trying to get a system problem resolved when no single vendor has systems responsibility can be excruciating. VARs are certainly aware of this pain point and have been trading on this “one throat to choke” benefit for a long time.
Anything new in this bundle?
So are integrated stacks a trend or a paradigm shift? Like fashion, technology also seems to run in cycles. From mainframes to open systems and back again, resources and compute infrastructures consolidate and decentralize on what seems like a continual basis, usually in response to changes in technology and/or new market needs. A good example is the movement back toward direct-attached storage to increase performance and the use of huge grids of white-box servers to assemble compute platforms at companies like Google. Rather than new technology, this integrated IT stack approach seems to be driven by a perceived need in the market for a simpler, large infrastructure solution; indeed, private clouds are a use case identified by these vendors.
Companies at the smaller end of the spectrum clearly appreciate bundles and plug-and- play solutions; they typically don’t have the people or expertise to make a complex system work. What may not be as obvious is the appeal of a simpler infrastructure in large companies as well. While they do have the talent to handle a sophisticated integration project, they’d rather have those people work on other projects, ones that can make them more competitive. So, reducing the complexity of large IT systems can be a big driver for them. This can be especially true given the growth of server virtualization, which is a big part of private cloud solutions and has added to the complexity of IT infrastructures in its own right. Integrated stacks, in part, are an attempt to address this need, plus provide better scaling and an improved support model. But there’s also something in this for the vendors, too.
What the vendors are getting out of this “movement” are new products to sell (bundling has always provided new price page items) and more control. With the commoditization of IT hardware, the established manufacturers have seen their sales numbers and profit margins squeezed by lower-cost storage products from new companies. And the constant improvements in CPU power that Intel keeps providing are generally shifting applications away from higher-end servers, reducing their considerable margins for the established companies that have provided them. While these developments may not have the big manufacturers anxiously eyeing the rearview mirrors for a wave of white boxes to overtake them, it certainly has their attention. Combining their existing products into a big-box solution is one way to help reclaim their stake in the enterprise data center. But there may be some problems with this strategy.
The difference between these integrated IT stacks and the solution selling that VARs have historically done is that these groups of “stack” vendors are leaving out a couple of key ingredients in their bundling strategies, namely choice and independence. They’ve not only chosen the products that will be included in the bundle, but also the configurations. While the result may not be a one-size-fits-all situation, it could force users into two or three sizes. Existing equipment represents another rub when trying to sell complete solutions from a single vendor; people like to know they can reuse gear they have on the floor, even if they choose not to. It seems that nobody, at least since the days of dot-com startups, is really interested in buying the whole soup-to-nuts infrastructure at one time. Throwing away systems in IT, especially storage, while they have some (perceived) useful life is very difficult--economically and politically.
But the real question is freedom. Do companies really want to put themselves into that (big) box with one vendor group? While having one throat to choke is appealing to customers, there can be vendor lock-in when the system bundle is supplied by the manufacturer rather than an independent VAR.
What’s a VAR to do?
Most of these solutions are aimed at the largest of the enterprise customers, and to date they haven’t created anything close to a groundswell of market acceptance. But if the strategy is successful in this space, it could surely be implemented in smaller bundles and sold to smaller customers. A few VARs may think integrated stacks are the beginning of the end for the VAR business, but most probably see it as just another battle to fight with their small armies. As it turns out, these issues of choice, existing equipment and independence may be all the ammunition a VAR needs to fight this new battle.
VARs have been using the one-throat-to-choke mantra as a sales strategy. But while offering up their throat, they can say with a straight face that the customer is still buying use-appropriate, cost-appropriate system components from different vendors. They can assure the customer they’re not being oversold a system made of best-of-breed components when a lower-cost solution would work fine for their environment. And, they can design a solution that uses some of the customer’s existing equipment. Perhaps, most importantly, a VAR can say something the integrated IT stack vendor cannot, when faced with a customer who’s concerned about putting all their eggs into one basket. They can say: “If you’re unhappy after we put in your system, fire us. You’ll have all the documentation, and since you’re buying independent components, you can run it yourself or hire another capable VAR to help you.”
Eric Slack is a senior analyst with Storage Switzerland.
This was first published in July 2011