The most notable shift over the years has been in the processor space: Single-core became dual-core became quad-core. This was by far the mightiest leap forward. What's surprising, though, is that it hasn't really translated into a surge in server sales. Sales have trended up, but there has been no great leap.
Users seem to expect vendors to bring out newer and better systems, and overall, they take product upgrades in their stride. It's almost as though they expect Intel and AMD to come out with something better, but they aren't inspired buy it any more.
What appears to be missing is a game-changing piece of disruptive hardware. If the ramp up to multi-core wasn't it, it's difficult to envision something else coming along that will do better in the market.
Of course, a look back through the past half-decade demonstrates that virtualization and blades have taken hold more and more each year and continue to do so. On the negative side, announcements of the fall of Intel and the rise of AMD were certainly premature.
So if nothing really earthshaking lies behind, what lies ahead? Here are a few predictions for 2009:
- SSD will gather steam
- The virtual honeymoon is over
- Doom and gloom
I can't wait to see the year-end stats for Solid State Drive (SSD) adoption. While they will remain a drop in the bucket compared to disk, that drop is getting bigger rapidly and it will soon be a perceivable puddle in the overall market.
If they can get the price down enough and knock out the few remaining technological kinks, this one is a stone cold certainty for rapid adoption: It removes the disk speed bottleneck by rocketing read and write performance, is far more reliable over the long term (eliminates moving parts), and consumes far less power. So the green IT lobby will demand it. But it's just got to become a must-have item as soon as it catches hold. People will pay $100 or $200 more to add SSDs when they are building their own machines at Dell or other sites. Then it's only a matter of time before it becomes standard. Maybe not in 2009. But by 2010, SSDs will be commonplace.
Virtual Honeymoon Over
I remember seeing a documentary about traffic in New York City in the early days of the automobile. Cars multiplied in unforeseen numbers, so they added bridges and roadways galore. The transportation commissioner in the 1920s commented that as soon as you built a new road, more cars appeared to take advantage of it, so you never actually made any real traffic progress in the long run.
It may well be the same thing with server virtualization. Now that the early adoption phase is over, we seem to be heading back to square one. Physical server sprawl has been replaced by virtual server sprawl. IT managers are losing track of how many virtual machines they put where and running what. Chaos has ensued in some shops. It's like we opened a 10-lane superhighway to unjam server deployment; now we find out that 15 lanes worth have actually been deployed, and the roads are jammed once again.
Doom and Gloom
It is impossible to escape the doom and gloom from the news media. Sadly, many valuable IT folks were among the record-breaking November 2008 layoffs. So there is no point in discounting the possible repercussions. Better get used to cutbacks, last-minute P.O. cancellations, budget cuts and more throughout the entirety of 2009.
To balance this out a little, note that it isn't all bad. P.O.s are still being approved. The recent IDC server stats from the third-quarter serve are a case in point. IDC had server revenue down by 5.2 percent year-over-year at $12.6 billion not good news, by any means. Yet unit server unit shipment growth was up 2.8 percent year-over-year. And while the low-end stats fell, high-end enterprise server revenue grew 4.0 percent year over year, the third consecutive quarter of growth for the segment.
Clearly, belts are tightening, but budgets are still finding room for hardware. Oddly enough, high-ticket items seem to be on the ascendancy.