Cloud Storage Infrastruggle and the Battle Against Marketecture
Develop more critical thinking of the product pitches you get every day from the growing number of cloud services vendors and you'll begin to see that the real issues are often in what's not mentioned in vendor marketing.
I suggested "The Infrastruggle" as the name for my column. I like that moniker because it summarizes in my mind what is going on right now in the IT industry generally. Vendors are trying every marketing trick they can muster to convince IT consumers to join their tribe -- a grouping of dedicated fans who, if they can be cajoled into deciding in favor of certain core technologies, will subsequently find their choices narrowed to a preferred subset of the universe of available options.
Buy Apple, and everything from your clock to your car will be chosen, at least in part, for its iPhone or iPad or iPod dock. Buy VMware, and you will shortly be considering a rip-and-replace of your storage area network (that really isn't a network) and its replacement by some "server-side software-defined storage array" (in other words, a direct-attached storage kit with some flash).
There is no infrastruggle if you just go with the flow, accepting without question the business value case and technical arguments for a given technology presented by your preferred vendor. Plus, there is certainly some comfort in wrapping yourself in all of that marketing sugar and fat, especially when the alternative is a hard slog against the tide and the very likely outcome that you will be characterized negatively as an iconoclast, or worse yet, as someone tragically un-hip. While the latter may appeal to your inner rebel, it won't get you a lot of invitations to vendor soirees or unexpected gifts delivered during the holidays.
If you start to develop a critical -- or, daresay, a skeptical -- attitude toward the products that are presented to you, asking what the vendors regard as "hard questions" about the truthfulness of claims about products and their capabilities, your morning e-mail will sow the seeds of day-long disgruntlement.
Even as I was writing this piece, I received an e-mail inviting me to read a report from EMC and Arrow regarding "factors forcing healthcare organizations to rethink their storage strategies." The report summary said that these factors were, in order of priority, data growth, inefficient use of storage capacity due to storage silos, changing data retention requirements, "and more…" While I fully concur with the notion that storage has become an issue for most organizations -- and that the growing data burgeon has something to do with that, as do the other two factors listed -- I have to wonder why we weren't pitched the two real causes for the interest in storage strategy.
For one thing, we are placing mostly unmanaged data onto mostly unmanaged storage platforms. This is a large part of what drives inefficiency and waste. Until we start practicing data classification, data hygiene and data archive, about 70 percent of our storage capacity will continue to be wasted and capacity requirements will continue to drive new acquisitions. Any strategic reconsideration of storage should start here.
Another huge driver of storage strategy reconsideration is simply cost. Depending on the analyst you read, storage now accounts for between 33 and 70 cents of every dollar spent on IT hardware. And while the EMCs and the Arrows don't want to raise the issue while trying to sell us a new kit, the fact is that IT budgets have not significantly increased even as the Great Recession recedes. The old Dorito's model of the storage vendor (remember the tag line: "Crunch all you want, we'll make more!") may be coated these days with comforting messages about product features that help us "do more with less." But, that "doing more with less" value is usually only realized by a significant new investment in new technology.
Truth is, from a total cost of ownership (TCO) perspective, an all-flash memory array may be more cost effective than an all-disk array from a power, HVAC and software licensing savings perspective, but those savings require a big purchase of flash technology, usually from a venture capital funded startup formed with a three-part business plan: create slide deck about product, get VC money, sell company to three-letter vendor. In the end, a lot of folks are going to find themselves stuck with flash products from companies that no longer exist if the oversaturated flash market implodes.
Truth is, cloud computing (or IaaS or software-defined data center or whatever they call it next week) is another name for outsourcing. We saw it in the 1980s during the Reagan Recession under the guise of Service Bureau Computing, and then again in the late 1990s dotcom meltdown in the form of Application Service Providers/Storage Service Providers (ASPs and SSPs). The name changes but the reality remains: When times get hard, a lot of companies try to outsource their IT, which after all "is not a core competency of our company." The resulting outsourcing arrangement tends to sour more quickly than a Kim Kardashian wedding (the first one).
Truth is, x86 hypervisor computing is not the only way -- or even necessarily the best way -- to realize greater cost efficiency from distributed IT. In many respects, VMware is already a legacy product: the iOS and Android model of delivering applications on demand, directly from storage, to the client device may well be the logical replacement for stacking up virtual machines on a physical host in a managed hosting facility ...er... "cloud" somewhere and waiting for someone to use the instantiated applications. No one talks about that today, but the contest between the two visions of server virtualization and application-on-demand goes back to the late 1990s when VMware was a smile on mommy's face that daddy didn't understand and Parallels was a direct competitor to VMware and hypervisor computing, rather than just a context switch for a Mac Book Pro.
The main problem with the hypervisor folks is a convenient deletion of history. VMware wants to sit atop the stack of "software-defined" IT infrastructure the way that IBM sat atop the stack of mainframe computing technology in the 1970s and 1980s. They seem to have forgotten what single vendor dominance meant from the standpoint of IT hardware lock-in and software costs (or maybe they haven't) as they pine for the orderliness of de facto standards (while breaking the de jure standards -- like ANSI T-10 SCSI standards a couple of years back -- that have been developed over the past couple of decades).
In short, the infrastruggle is, in many respects, the battle to shape our memory of the history of IT and the lessons we draw from it. Our choice is simple: Pay attention to the realities of cost-containment, continuity, compliance and carbon footprint reduction and make strategic decisions to address our business requirements within sensible boundary conditions, or go with the flow and participate in IT's race to the bottom. Apologies in advance if this column is not filled with the kind of sugar-and-fat marketecture that you are seeking, but if you find it provocative and interesting, welcome.
Jon Toigo is a 30-year veteran of IT, and the Managing Partner of Toigo Partners International, an IT industry watchdog and consumer advocacy. He is also the chairman of the Data Management Institute, which focuses on the development of data management as a professional discipline. Toigo has written 15 books on business and IT and published more than 3,000 articles in the technology trade press. He is currently working on several book projects, including The Infrastruggle (for which this blog is named) which he is developing as a blook.