Repetition, The Refuge of Flawed Ideas

I just finished reading a column by a guy who works for a second-tier IT "research and analysis" firm who claimed that a growing number of small firms were adopting "SANs," by which I assume he was referring to Fibre Channel fabrics. To find out, I would have to register with the Website of the fellow's publisher, which I have no intention of doing, mainly because the site publishes this fellow's nonsense. Unfortunately, someone emailed me the entire text of the article, probably just to irritate me.

I can think of no more patently absurd propaganda flying around the storage market today than the idea of storage area networks (SANs) for small to medium enterprises (SMEs). SMEs have no requirements whatsoever that necessitate the acquisition and deployment of an FC fabric. The notion that SAN adoption is climbing among SMEs flies in the face of reason, and no matter how many times someone in the analyst community is paid by a vendor to say it, it is still untrue.

In this case, the analyst based his case on an interpretation of statistics drawn from 220 companies, 10 percent of which were firms with over 500 employees, 34 percent in the 100- to 499-employee range, and 56 percent in the sub-100 employee rank. Apparently, the guy associates "big" and "small" with numbers of employees; no mention is made of earnings.

He seemed sufficiently energized by his less-than-scientific-poll to assert that SANs are on the upswing: 35 percent of respondents said they have deployed "networked storage." But no definition is offered regarding what constitutes networked storage--does that amorphous category include file servers with direct attached arrays? What about my laptop hard-disk, which is accessible to my desktop when I turn on my wireless LAN connection? Another 13 percent are "actively evaluating" networked storage solutions and another 26 percent are just "considering" it. And 2 percent, bless them, are saying that, like me, they aren't really sure what networked storage is exactly.

In real numbers, that means that 77 companies had some sort of storage that they viewed as "networked," possibly nothing more than an NFS volume accessed across a LAN. Another 86 firms were "thinking" about networked storage (don't we all), four companies didn't know what networked storage was (most of us don't), and 53 summarily rejected the entire notion of networked storage outright as inappropriate to their requirements.Since there was no previous study with which to compare these results, I fail to see how any trend could be discerned from the data. That makes the entire premise of this shoddy piece of "analysis" a bit inane, if you ask me. It must have been a very slow news day indeed for a major storage pub to run this drivel.

Excepting for a very small percentage of SMEs that leverage niche applications like medical imaging, geographical information systems (GIS), audio/video production and post-production, database sales, on-demand video, or computer-aided design (CAD) systems, most SMEs use applications like Exchange Mail, SQL Server, Office 2003, ACT!, QuickBooks, and maybe the occasional specialty app developed to support their primary line of business: retail sales, manufacturing, financial, restaurant or other services. If these companies have more than eight to 10 commodity servers, it is unusual. If all of these servers mount the latest operating system from Microsoft, it is remarkable. And if their entire IT department of a smaller firm is larger than one guy named "Roy," I'd be surprised. In most small firms, there is no IT department at all.

The average data complement in the SMEs that I have visited ranges from about one-third of a terabyte to 3 or 4 Tbytes for the largest players in this class. Typically, if the entire storage infrastructure in these firms exceeds $50,000 in total cost, it is anomalous. Just try selling them a $19,000 array. If you don't believe me, a lot of Serial ATA array vendors are going broke trying to do just that. Worse yet, try to sell them a FC fabric at $2,000 to 3,000 per port, which is somewhat greater than their entire investment in application servers. They will laugh you out of the office.

For some reason, the analyst pursued his flawed premise with a data protection bent. SMEs were flocking to networked storage for disaster recovery. Not so subtly, he was trying to argue that SME SAN adoption would parallel early fabric adoption patterns in large enterprises, when the topology was viewed as expedient for consolidating backup processes and sharing expensive backup systems. That's where he undoes himself.

SANs are about as useful to the SMEs as high-end tape silos. A friend of mine in the tape backup racket just returned from Russia, where he was brought to a government installation to address what was described by his local representative as a "huge tape opportunity" for his company. After making a technology presentation to a large and enthusiastic audience, he was escorted down to the vaults where he saw the target: lots and lots of round tape reels, maybe a mile of them.Doing the math, however, he discovered that converting the entire complement of data stored on those backups to present day tape technology and formats would require exactly two tape cartridges; four, if you wanted a redundant set. It was hardly the kind of multi-million dollar deal that my friend was used to closing, and it bears a tremendous similarity to the SME opportunity for SAN vendors who want to peddle it as a backup solution.

I suppose to sound authoritative and consultative, the fellow switched gears at the end of the piece and started issuing motherhood considerations for SMEs considering SANs: buy carefully, consider all of the hidden costs, get your data organized. In effect, he was beginning to list all of the reasons not to buy a SAN if you are an SME, which is what he should have started with in the first place. Instead, he rounded out with the assertion that SANs are no longer "the exclusive domain of the digital elite," bolstering the view of SANs as something sexy or exclusive, which they certainly aren't.

Vendors pay big money to analysts every year (this guy's company charges you $10,000+ just to be on their radar) in what amounts to a love-hate relationship. Depending on who you ask, the analysts are the paid political spokespersons for the industry, or they are running an extortion scam on their vendor clients worthy of a federal racketeering case. What they specialize in, when it comes down to SME SANs, is articulating a fantasy that the vendors hope, if repeated often enough, will become self-fulfilling.

The subsequent article illustrates the problem in spades. It is exactly the type of analyst crap that is causing all of us to get cynical about the value of our self-appointed advisors and oracles.

Do yourself a favor: When you read something by one of these guys that makes no sense, make a note of the name. The next time you run across a web site or magazine that is publishing the propaganda, click off the site or throw the magazine into the circular file. That's what I will be doing with this guy's ruminations going forward.Some audacious analyst claim got you pitching in to your own circular file? Drop Toigo a line