As 2002 Goes Into the History Books, Sanity Rules
We've pulled together some of the noteworthy moments for enterprise storage in 2002 and look ahead to what this year will hold.
At the Computer Measurement Group (CMG) annual conference in Reno, NV a few weeks ago, I had the opportunity to deliver two presentations and sit on a couple of panel discussions about the future of storage. I’d like to believe that the audience was filled with exactly the people who read this column: practitioners in the arts of systems and storage design, deployment, management, and measurement. By that I mean that the audience was a tough crowd and unwilling to listen for very long to hype and "marketecture." even from Gary Bloom, CEO of Veritas Software, who provided the keynote on the opening day of the conference.
As Bloom's presentation—more an infomercial for his company than the presentation of a thoughtful engineering paper—wore on, many CMG’ers did what CMG’ers are likely to do when confronted by a talk short on technology and long on "marketecture." They walked out. It was a microcosm of how the world at large is looking at the storage vendor community right now.
I forced myself to listen to Mr. Bloom’s talk from start to finish because it provided a good wrap-up of 2002—the year that Storage Management issues took center stage away from storage hardware. Bloom filled his presentation foils with lots of Gartner Group charts showing the importance of storage management, including one that suggested that the 13 percent of the budget companies spend on management software determines the cost efficiency of the 29 percent of their budget spent on hardware and the 58 percent they spend on labor. Gartner is good at putting together scientific-sounding numbers to illustrate the intuitively obvious and their 13-to-29-to-58 ratio was probably about right.
Bloom went on to say that virtualization (providing applications with scaleable storage) and backup remained the key pain points for organizations today—familiar themes to readers of this column—but he surprised me by talking about virtualization on the switch. The switch controlling access to storage devices organized in a “pool” would provide a service that would create logical volumes from physical storage, so you could forget virtualization plays by out-of-band storage server companies such as StoreAge, or in-band virtualization appliances from DataCore Software or Fujitsu Softek, or array controller-based solutions from EMC, Hitachi Data Systems, IBM, et al.
Though he didn’t come out and say it, you could also forget host-based virtualization schemes like those championed by Veritas! Interesting point, that, and perhaps a signal that Veritas was doing a deal with Cisco or one of the many Fibre Channel switch vendors to move their Volume Manager off the application host and onto a switch. I will monitor this closely and you should too.
Bloom also took the opportunity to diss the “grid computing” concepts being developed by many hardware vendors presently, disparaging them as more hype than hyperclusters, and said that the real work to be done was to make clustering software more application aware—that is, able to bring up redundant storage, network and CPU resources when an application’s performance dips below some pre-established threshold.
In short, Bloom provided the party line of the industry, with a generous dollop of Veritas-speak, to a crowd that was mostly tired of hearing what vendors were saying. To me, 2002’s most important trend was exactly that: it was the year that IT pros stopped believing the hype and began demanding real solutions with measurable performance. David Wright, CEO of Legato, made that observation in Germany last summer: consumers don’t care as much about extra bells and whistles these days; they want products that execute and deliver what they promise. Time has proven him correct.
As this new sanity carries into 2003, it may also provide a springboard for new purchasing patterns. International Data Corporation just released a highly contentious report about a dip in overall storage acquisitions in 2002. What made the report contentious was simply that IDC placed EMC third in market share, behind HP and IBM. The boys in Hopkinton were sorely disappointed by the report from the neighboring analysts in Framingham, to whom they had paid a lot of money over the years. They responded with reports of petabytes of storage sold in the previous quarters that, they argued, clearly put them ahead of the pack.
They could have kept mum, because no one cares who ships the most boxes anyway. What will matter in 2003 is if many boxes are shipped back to their vendors. According to a money analyst at Credit Suisse Bank, the three-year leases on equipment shipped in 1999 (the last boom year in storage) are coming due as 2003 approaches, and the preponderance of consumers she had interviewed were sending the equipment back rather than renewing their lease. The reason: you can buy a lot more storage hardware for a lot less money today than was predicted by the “storage on demand” programs of the late 1990s.
This may be the real cause of consternation for EMC, IBM, Network Appliance, and the rest, but it is also cause for celebration in many smaller-storage-companies-you-never-heard-of, particularly firms that sell used gear or storage devices that are plug-and-pin compatible with brand name vendor wares. These vendors stand to make big bucks giving end users what they need at a reasonable price. Many companies report that sales gathered steam in late 2002 and show no signs of slowing down.
Meanwhile, IDC also reported that it is scaling back its revenue projections for NAS and SAN vendors, mainly because networked storage has not caught on in the market as IDC had originally projected. People are trying to find ways to make more use of the storage they have through better management and they prefer to buy cheap server-attached storage as a hedge against the deployment and management costs of new-fangled networked storage topologies. Before we go thanking IDC for this sane commentary, those of us with the ability to remember things that were said more than three minutes ago should recall that IDC was instrumental in stirring up the SAN hype in the first place.
Hopefully, 2003 will be the year that sanity rules in your shop.