In-Depth

Bringing Order to the Storage Universe?

Our recent column about historical revisionism and EMC brings immediate reader response.

• By Jon William Toigo
• 02/26/2004

Few columns have touched as many raw nerves as our recent piece on historical revisionism and EMC Corporation. The piece recounted a meeting with an EMC insider who endeavored to correct my interpretation of events surrounding the creation of the Storage Networking Industry Association (SNIA) and its Storage Management Initiative Specification (SMI-S) and the behind-the-scenes role played by EMC—or rather the perceived threat posed by EMC.

In politics, perception is everything, and the perception of a burgeoning threat from EMC to dominate the storage industry (a la Microsoft in the server world) and to impose its own proprietary protocols and architecture as de facto industry standards seemed to have been a key driver behind the industry’s embracing of SNIA, then SMI-S.

My source from Hopkinton did not challenge this view, but sought to modify one element: EMC was not really trying to dominate the industry, he argued, but to foster industry cooperation by threatening to dominate.

It was an interesting backstory on the creation of SNIA and SMI-S and seemed worth reporting. At the same time, we sought to provide context with the observation that EMC had been on a buying spree of late and that its purchases added up to a one-stop-shop solution set that made its old proprietary ways seem like child’s play. In addition to buying its own operating system, file system, hierarchical storage management scheme, and data tagging controller technology (used in its Centera offering), the company seemed to be on track to purchase a tape vendor. That way, they would be able to offer a soup-to-nuts solution to consumers who have grown tired of serving as “the point of integration” for the anarchical world of storage industry vendors.

The first comments came from readers as far away as Belgium and as nearby as Irvine, California. Seemed that everyone wanted to know which tape vendor EMC was going to buy. One e-mail stated that ADIC was the rumored target six months ago, but StorageTek had recently surfaced as another “probable candidate.”

For the record, we have absolutely no idea what EMC’s acquisition plans are, nor do we have any inside scoop on acquisition targets.

A reader who works for a leading tape technology vendor contributed this observation: “I found it curious that you should speculate that EMC might buy a tape library manufacturer. While EMC's [Information Lifecycle Management] Initiative has caused it to move away from its past "tape is dead" stance, I don't look for EMC to buy a library systems vendor (any more than I would expect EMC to buy a tape drive company) simply because they can OEM the tape drives and library systems they require, just like the other platform OEMs (other than IBM—and IBM buys mid-range library systems from ADIC, and tape drives from just about everyone). EMC can sell a ‘one-stop shop’ ILM solution and still OEM their tape drives and library systems.”

The comment is well taken and we do not disagree, except that EMC has already made purchases that could instead have fashioned as OEM agreements between Hopkinton and the companies it has absorbed. Someone there obviously likes to own the entire supply chain and is not content with OEM agreements.

From a noted analyst and editor, another interesting perspective: “My take on WideSky is slightly different: I think it gives EMC a major differentiator in that EMC can embrace all the SMI-S stuff as amorously as it wants—(Geoff Norman's first law of IT standards is that a vendor's interest in standards is inversely proportional to its market share)—while offering the strongest story on incorporating legacy boxes into a standards-tending environment. Standards are fine for users with new or recent kit that conforms but—as the manufacturers know, and as Microsoft has raised to an art form by basing its hegemony on backward-incompatible file formats—the extra benefit for vendors is that customers are obliged to make any ageing kit obsolete before its time.

“Your mention of [former EMC CTO] Jim Rothnie reminds me of an interesting incident at an analysts' get-together in London in the autumn of 2000. I had a bit of a rant about the crippling effect that mutually exclusive management APIs were having in user-perceived function, and the onus placed on third-party software developers to build and test for every API. Jim grabbed me by the lapel and dragged me out into the hotel corridor, where a heated, finger-wagging row ensued, leaving us both unconvinced of the other's case. Many e-mails on the subject crossed the Atlantic in the following months. To be fair to EMC analyst relations, about 18 months ago, when EMC was busily espousing the cause of industry-wide standards, I received an e-mail with the simple words—‘You told us so.’”

The final word, for now, comes from a reseller who e-mailed to advise us, after reading the column, that whatever strategy EMC was pursuing, they had better do it fast.

“We are in the process of scrapping out the three- year-old Symmetrix platforms that [our client] paid [$]6 million for in 2001. Here is what we have found so far: There ain't much value in this equipment. Sym Frames, without drives, are bringing $15.80 each at the scrap dealer. What does that mean from an ROI perspective?”

What indeed? Our take on all of this is simple: the industry continues to wring its hands over what EMC is going to do next. Do the company’s recent highly publicized agreements with Outer Bay Technologies in Campbell, CA, and with Oracle directly, to find ways to carve reference data out of databases, further bolster its claims to be a purveyor of a true ILM solution? Or does it require a suspension of disbelief to hold out hope that someone will bring order to the storage universe?

Watch this space.