Microserver Wars and the Threats Facing Intel

Rob Enderle
Slide Show

Top Five Server Monitoring Battles - and How You Can Win Them

Well, it took a few days, but after HP announced with Project Moonshot that it will be coming out with a Calxeda ARM-based hyperscale server, SeaMicro, which sells servers - some of which are remarketed by Dell - emailed out its response (which is similar to Intel's response). Unfortunately, HP wasn't the only company announcing an ARM effort.


The SeaMicro and Intel responses, as you would expect, were focused on the fact that they have been in this microserver market for some time and pointed out that HP had basically launched a science experiment in that actual hardware that could be purchased was a year out. Microservers are expected to be the fastest-growing server segment this decade and since they are specifically targeted at cloud loads, they could become the dominant server type by decade's end.


The battle behind the scenes isn't between SeaMicro, a small, young specialty server maker, and HP, the largest server maker in the world, but between ARM, which is attempting to move up market and Intel, which is attempting to move down market.


What I mean is that ARM is attempting to move from a low-margin business based largely on smartphones and, more recently, tablets into servers and PCs, while Intel is attempting to move in the opposite direction. So far, ARM, which is made up of a consortium of competing companies, is having more success and that is because it is vastly easier to move up market than down market for any vendor.


Let's explore both fights.


Automotive Example


Often, we use the automotive market, which has been around much longer than the tech market to showcase differences. We've seen attempts from premium vendors to move down market and value vendors moving up market. The up-market moves were exemplified by Lexus (from Toyota), Infiniti (from Nissan) and Acura (from Honda). Down market was Cimarron (from Cadillac) and Versailles (from Lincoln), both of which failed. BMW did move down market with the Mini, showcasing that it can be done, but I think you could argue that the Mini, while successful, isn't in the same league with Lexus or Infiniti going the other way.


This is because moving up market from thin margins to thicker margins leverages the cost efficiencies of the low-cost vendor and is additive. The more successful it is, the more profit the company makes. However, in moving down market, a premium vendor must reduce costs to compete and often isn't as efficiently structured. The vendor actually feels the need to remove its premium features and the result is either a crippled or damaged product set, which defined both the Cimarron and Versailles car efforts. What made the Mini work is BMW actually created a bottoms-up new company to address the market. It didn't try to produce de-featured BMWs.


In short, if Lexus steals a Toyota customer, Toyota is fine with that, actually ecstatic, and if Mini steals a BMW customer, BMW isn't happy, but it can't be prevented because the two entities are separate. BMW recognizes that it might have lost the customer anyway and that it is better to make a sale than lose the customer entirely. In a typical down market effort, the down market product is crippled so that no one would reasonably pick the down market product to protect the high-end offering. While the vendor thinks the choice is between its two products, it is actually between its product and a third-party offering, and this decision generally gives the business to the third party.


I think the BMW approach points the way for Intel and this isn't the path it is currently on.


x86 vs. ARM


While Intel would appear to be at an organizational disadvantage, the ARM attack isn't from one of the tier-one ARM vendors, Qualcomm or NVIDIA yet and only the HP brand is making it real at this point. HP is a company with multiple wars and Intel is a powerful partner. With this battle, played at an organizational level, Intel should be able to get HP to put more resources on Intel's offerings over its ARM alternatives without arm twisting because the Intel relationship is more important to HP. However, should NVIDIA or Qualcomm enter this fight or should AMD, which is in the process of reorganizing for a reason, shift to this opportunity (AMD and HP have historically been very close), this dynamic could quickly change. As it exists today, this is still Intel's market to lose.


Wrapping Up


Wars are often lost by political decisions made internally (going back thousands of years). Intel faces one of the greatest threats it has ever faced from ARM and has so far been unable to take the battle into ARM's smartphone, MP3 player and tablet territory, while ARM is increasingly encroaching on Intel's space. This is because Intel isn't resourcing the battle effectively and likely because it is crippling its low-cost efforts to protect its margins (a common complaint from its OEM partners). ARM could be doing better as well if a tier-one ARM vendor were focused on the server space or the ARM effort were more consolidated and focused in general.


In short, this microserver market is still Intel's market to lose and it is the only technology widely used in it. However, this will change dramatically over the next 24 months and if the ARM attack can be better resourced and focused before Intel can deliver a more competitive offering, we'll have another example of a bad way to move down market and one more instance where a defender was eliminated as a result of its own bad choices.


Microservers are the growth market for servers and whoever gets this market will likely eventually be the new server king. Typically, when a market turns, the dominant vendor doesn't make the turn because it protects its existing products too long. SeaMicro, the vendor that led this Intel defense, in its email to analysts also said "our technology works on ARM equally well," which, I think, showcases the risk.

Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.


Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.




Subscribe Daily Edge Newsletters

Sign up now and get the best business technology insights direct to your inbox.

Subscribe Daily Edge Newsletters

Sign up now and get the best business technology insights direct to your inbox.