The overwhelming majority of sales in the high-end storage server market is distributed among three players: EMC, Hitachi and IBM. Talk to the manager of any large enterprise IT shop and you are likely to hear strong opinions about all three companies, their products, their service and just about anything else that may be even remotely connected to their technologies and business practices. And why not? Any IT manager who has invested in one of these companies' high-end products has committed a lot of money and resource towards the purchase, care and feeding of these behemoths.
And because of the sizeable investments that are required for EMC's Symmetrix, Hitachi's Lightning and IBM's Shark platforms, IT managers often bet more than their budget on these beasts - they also may be betting their careers.
During the last few years, particularly as the robustness of midrange devices has increased dramatically, there has been a good bit of speculation that these largest servers would go the way of the dinosaur, rushing towards extinction. In a sense, that is what is happening.
I once worked for a company that was doing its level best to reduce costs by eliminating mainframe servers. The running joke was that these were dinosaurs living in our glass house, and no day passed without someone making a comment that this was the day we were going to stop feeding the beasts and just step back and shoot them.
Mostly they ran Cobol applications, and as we were all signed on to the concept of structured programming (which Cobol clearly wasn't), we anticipated that the impending death of Cobol would drag down the mainframes as well.
What none of us realized of course was that the sheer volume of Cobol code out there resulted in the language reaching critical mass by the late 1970s, at which point, Cobol had become sense self-sustaining. Cobol, we eventually learned, was never going to go away. And the extinction of those mainframe dinosaurs, like the death of Mark Twain, was more than a trifle exaggerated.
What happened with the mainframes was that the big iron simply evolved. Management had gotten used to the functionality they could provide, and couldn't be pried away from them. Some of them even morphed into Linux boxes.
Something similar is happening to the big storage servers as well.
Lots of us have watched with interest during the last several months as the high-end providers reconfigure themselves for the new century's marketplaces. EMC and IBM are emphasizing the software side of their houses, EMC with its Widesky and IBM with its Tivoli product set.
Hitachi has positioned itself for the midrange by opening up its API to Hewlett-Packard, and sells into the midrange by using HP and Sun as channel partners. EMC has brought the Clariion line back to life with a vengeance, and through its partnership with Dell will use it to attack the midrange.
But none of us should expect any of these players to walk away from their high-end business, which in most cases is a moneymaking machine. The boxes will certainly change, and more high-end features will be passed down the line to the midrange, but the big iron - or at least the features of the big iron - will be with us for a long time to come.
And that is probably all to the good because even though these beasts will most likely represent a decreasing slice of the overall storage pie in the coming years, they still are probably going to be the proving grounds for many of the best of the bulletproof features that lots of you need to see.
Individual functions and capabilities will continue to migrate to the mid-range, just as hierarchical storage management and RAID did in the past, and even those of us who never spent a nickel on these machines will eventually have access to some of their features.
So keep your fingers crossed that high-end storage carries forward for many years to come. You may not want one in your own shop, but you can be certain that many of the features developed for them will move down the food chain into the price range you can afford. Then maybe an IT manager can buy some of those high-end capabilities without having to bet the farm on the purchase.