In-Depth

Spring Busts Out: Part II

From "thin provisioning" to a report on the storage outlook in the small to medium-sized enterprise world, we look at the silly side of Spring.

• By Jon William Toigo
• 04/14/2005

With the flowers come the weeds. And this Spring storage season is showing no shortage of the pesky things. Here are a few examples.

The vocabulary of vendor-speak is changing once again. Having failed to communicate the value of virtualization—or, more accurately, having poisoned the value proposition through incessant vendor infighting and campaigns of fear, uncertainty and doubt—many vendors have now adopted the colloquialism “thin provisioning” as a substitute. Thin provisioning comes from the dictionary of utility computing, where it basically means fine tuning services or resources in response to changing workload requirements.

I seem to recall that this is what virtualization was all about from the get-go. You take a lot of things (disks, LUNs) and represent them as a fewer things (storage pools) for simplicity’s sake. Why? Simple: managing fewer things is easier; growing the capacity of a virtual thing is easier; tuning virtual things to meet changing requirements is easier.

So why not just stick with the term instead of substituting a euphemism? Aside from the marketing stigma introduced by competitor and analyst hype, virtualization is terrifying to most brand-name vendors. Virtualization ultimately means usurping the functions currently built onto array controllers and reducing all arrays, regardless of their pricey branding, to the status of overpriced JBODs. With true virtualization, even RAID goes away, replaced by automated copy-on-write functions managed by the virtualization engine. Needless to say, the Big Iron array vendors are going to fight the idea tooth and nail until either the Virtualizers or the Big Iron mavens bite the dust.

Bottom line: Thin provisioning is just the latest in what will likely become a litany of colorful expressions for exactly the same thing: virtualization. The downside is that (1) the constantly shifting terminology may continue to obfuscate a clear-headed discussion of virtualization, and (2) we will have to do a lot more deciphering every time we sit through a vendor pitch.

The proliferation of products aimed at small- to medium-sized enterprises (SMEs) is occurring daily. Everybody from the Big Iron brand names to the up-and-coming newbies are contending for SME mindshare because it is still the only market segment showing any real growth.

A bit of background: Wall Street is predicting that the size of the large enterprise market is actually shrinking (more acquisitions, fewer customers) and that enhancements to their platforms by the three-letter companies is driving out product differentiation. The result is clear: there will shortly be fewer than three points of market share to separate IBM, EMC, and HDS in the top-tier enterprise market.

Most vendors are, therefore, hunting up new revenues in the SME space, though most are clueless when it comes to SMEs. They do not understand the world as seen by the small guy and are working to palm off dumbed-down versions of their enterprise products on the sub-Fortune 500s.

For example, EMC just paid IDC to write a profoundly silly report on the storage outlook in the SME world. (If you want a good laugh, download it at http://www.emc.com/analyst/pdf/EMC_Storage_White_Paper_030805.pdf.) One reader of this column sent me an e-mail to bring the document to my attention: “I read the IDC/EMC whitepaper in the bathroom today. You know it was a lot like reading People magazine: it took some time and kept my mind busy … but I don't think I learned anything. Do people actually get paid to write this worthless drivel?” The guy is an integrator, by the way, who specializes in providing high-availability storage solutions mainly to SMEs. He knows whereof he speaks.

Expect to see a lot of this kind of thing—sponsored endorsements and such—cropping up as name-brand vendors work feverishly to convince smaller firms that they should engage in the time-honored practice of “oversubscription with underutilization” that has made the vendors big money from the Fortune 500. Just don’t believe any of what you read.

Trust me: you are not lagging behind the technology curve if you don’t deploy a Fibre Channel fabric, and you should never purchase more technology than you need on the premise that it is strategic and you will grow into it. It is best to understand exactly what your applications require and to purpose-build your storage to meet those requirements. The most strategic thing you can do is invest in a good management software package, then buy only gear that can be managed via the software. (Unsolicited plug: Tek-Tools’ Storage Profiler is gaining big market share in the SME space because it is hassle-free and really works; and we are soon to test CommVault QiNetix in our labs.)

Without meaning to pick on EMC in particular, I have to note another annoying disruption in my Spring-time reverie: all the foo about enterprise content management and Documentum.

Here’s the background as I see it: Having waged an unsuccessful effort to push WideSky down the throats of the storage industry, followed by an equally unsuccessful effort to deliver the full set of functionality promised in its Enterprise Control Center (ECC) management suite, EMC last year changed its messaging to promote something called Information Lifecycle Management or ILM. They would talk incessantly about ILM this, ILM that, but when I asked them directly what they meant by ILM, they responded, “Well, what do you think that it means?” The conversation went downhill from there.

Now, they are pushing Documentum, their recently acquired content management application, as a cornerstone of ILM. With Documentum, you can classify data at its source so that it is easer to migrate to various flavors of storage (all EMC boxes, of course) later on. This is true enough. Any content management system would give you the same capability.

The problem, though, is that you generally need to rip and replace all kinds of applications you already have in order to make content management systems work. Are you ready to force users to break with their current apps and to work through the Documentum modality? This is a classic example of changing your business processes and workflows to fit one vendor’s idea of perfection rather than adapting a vendor’s technology to fit the way you do business.

While there may be business processes in your company that could be made more efficient by imposing the order and discipline of a content management system, I really have to wonder how many businesses have reached the level of pain that would make them want to endure a complete re-engineering effort of the type implicit in a Documentum implementation. In my experience, there are numerous fixes and workarounds for information management inefficiencies (and most are less expensive and less disruptive than migrating to Documentum) that companies will want to try before they sell their collective souls to Hopkinton.

Then again, EMC has made its fortunes selling refrigerators to Eskimos, so they are bound to sell a few copies of these wares. I expect a lot of noise every time they do.

I just wish they’d ratchet back the noise a few decibels. It hurts my ears. So much for Spring flowers and weeds. I welcome your comments at jtoigo@toigopartners.com.