There's no denying server blades are on the fast track. Blade sales topped $583 million in 2003, according to the latest edition of IDC's Quarterly Server Tracker. While this represents a small piece of the $13.7 billion 2003 server pie, it is a huge leap from 2002's $90 million in blade sales.
At their most simple, blades are thin servers that cram stackable computing power into a low power, heat-optimized sheath-like chassis. Ever since RLX unleashed the first blades on the market in May 2001, heat control and space savings has been the segment's main selling point.
The blades market underwent a tremendous growth spurt in 2003 that left the leading players jockeying for the top spots: After nearly two years of controlling the market, HP was elbowed down to second, and Big Blue came in first in terms of both units shipped and revenue. IBM, which had slightly more than one year of air time in the blade market, credited its success to the fact that it was able to leverage its mainframe background into the slim 2-way and 4-way Xeon blades.
Later this week, IBM will introduce a new BladeCenter-T design ruggedized for military and space travel type applications. The BladeCenter-T will be tailored to the telecom industry's Network Equipment Building System 3 (NEBS 3) requirements, and it reinforces IBM's push toward the high end. "These systems have been hardened to withstand high temperature, violent shaking, lightning strikes, airborne contaminants, fires, and electrostatic discharge," IBM spokesperson Chris Rubsamen told ServerWatch.
But blade vendors need not limit themselves to the high end, as evidenced by the strategy of vendors currently offering Pentium III processors. "No one's trying to be everything to everybody, which makes sense," Gartner Research Director Jane Wright told ServerWatch. Wright says vendors offering Pentium III or Pentium Mobile blades aim primarily for edge or high-performance computing environments (i.e., segments where distributing the workload to many not-so-powerful servers quickly is of value), whereas those vendors that add the latest and greatest Xeon servers to their fold are aiming at the high-end, mission-critical market.
HP's Adaptive Enterprise covers both bases with an eye on blade ubiquity. "HP sees blades as systems, not servers, existing within a larger context, persisting beyond a single generation of technology," said Anthony Dina, manager of business development for blades in HP's Industry Standard Servers group. "They're designed to help in every role where computation work needs to get done, desktop and server alike," Dina added. HP's ProLiant p-Class aims at the high end, with 2-way and 4-way Xeon blades in the stable, and an Opteron server in the pipeline. HP's e-Class is targeted at low-end edge apps.
Dell, on the other hand, sees its sweet spot at the low end. Despite a lack of emphasis on its blades, there was no keeping the Poweredge 1655 MC from the mainstream. Based on both IDC's and Gartner's accounting, Dell's blades took third place (albeit a very distant third place) in revenue. "We continue to closely evaluate the server blade market and gauge readiness for a second-generation server blade," a Dell spokesperson said. The Dell spokesperson with whom we spoke declined to be interviewed in depth, and provided no road map details. He concurred, however, with IDC's view that "2004 will be the beginning of a trend toward more significant blade adoption by customers."
Although neither RLX nor Egenera matches Dell in sales, both vendors have significant roles in the space, not the least of which is their first to market positions. The vendors have a year-plus head start over HP in blade development as well as blade-centric management software that each believes gives it a leg up. The blade market is currently one of the few hardware niches where innovative, smaller vendors can take a crack against market consolidation.
Even the research firms are taking notice of this. "We're really excited at Gartner because we haven't seen this in ages," Wright told ServerWatch, "[In other segments] All we see is vendors going away, merging, and being eaten up, and it gets a little depressing." Wright cites the exciting possibility that "someone like an Egenera or an RLX can consider being a major player here."