Gartner put out a five-year outlook for blades a few weeks ago. While finding that blades are the fastest growing segment of the server market, and they now account for 10 percent of all server shipments, Gartner predicts they will account for only 19 percent of server shipments half a decade into the future. I beg to differ.
Hard-Core Hardware: Drew Robb explains why Gartner is wrong about where blades are going.
Gartner’s logic is that blades are not yet mature, so there are still plenty of upheavals lying ahead. The report zeros in, for example, on a lack of chassis standardization as a major issue.
“Blade servers have a number of proprietary aspects, and the market lacks interoperability standards that lock users in to the technology,” said Gartner’s John Enck. “In today’s technology, the chassis represents an infrastructure boundary; compute, storage and network resources cannot span multiple chassis.”
Recent Hard-Core Hardware
» IBM Server Snapshot: Big Changes at Big Blue
» Recession, What Recession?
» Solid State Drives: No Flash in the Pan
But is this a good enough reason that blade growth should effectively be the same over the next five years as the last five? Absolutely not. Gartner thinks 19 percent of server shipments will be blades by the end of 2012. My guess is at least 30 percent for a variety of reasons.
Blades are sexy, towers are drab. Blades are cool, rack servers are relatively boring, at least from the perspective of the vendors, analysts and industry press. Users are going to be hearing a LOT more about blades than any other form factor. Even in those holdout shops that are quite happy with racks, it’s a cinch that some executive is going to insist on blades sooner or later.
Vendors like IBM and HP are headlining their green efforts with an emphasis on blades. They see it as cost-effective to roll out energy saving features in an integrated cooling and power package as well as a smaller form factor.
There seems to be an inevitable transition ongoing from direct-attached storage to network attached. It’s happened with the SAN. It’s even happening with the consumer sector where people store their e-mail online for free in some vast repository. More and more consumer data now resides in the cloud rather than on the desktop. So it makes sense to eliminate the hard drives once and for all and hook a bank of blades up to a central storage pool. As 8 Gbps Fibre Channel and 10 Gbps Ethernet catch hold over the next couple of years, this will make even more sense.
4. Thin Client
Thin client is going to get very very big over the next few years. Traditional desktops are becoming dinosaurs that are likely to be refreshed out in favor of thin clients that hook up to blades on the back end. This is currently in the early adopter stage but once it catches hold, I expect this to literally sweep the land.
Unsure About an Acronym or Term?
The old ENIAC filled up a massive room. Steadily over the past years, electronics have gotten smaller. Remember that old TV set from the fifties. Instead of going about four feet back, it has shrunk to only about two inches. How about music — that vast pile of albums that took up an entire wall in your bedroom now don’t take up more than a fraction of the latest iPod. The blade, then, just represents the inevitable transition of servers down to smaller and smaller form factors. People want smaller. Down the ages, we’ve gone from messages delivered by real live persons, to letters delivered by horse, to post office delivery to e-mail and text. The trend is always less space and smaller form factors. If you can pack the same computer power into a few racks that used to take up a large room, why wouldn’t you — provided you get your act together on power and cooling.
Of course, not all workloads play well on blades. And many organizations face a nightmare scenario in facilities with grossly inadequate power and cooling infrastructures. These factors will temper out and out blade dominance, but only a little.
Current stats support the above. IDC reported 40.8 percent growth in the second quarter for blades.
There may be one way, though, that the Gartner prediction comes true. Suppose blades get supplanted by an even smaller form factor — the “dagger,” the “pen knife” or perhaps the “nail clipper server.” After all, pizza boxes had about six months in the limelight before blades overwhelmed them. The same could happen to the blade. The small will prevail.