By any measure, microservers are the right solution at the right time. Just as the enterprise is starting to contemplate scale-out infrastructure for its own Big Data-facing private clouds, the microserver provides a low-cost, compact, highly parallel architecture that can quickly adapt to changing workloads.
But if that is the case, why has the technology failed to emerge as the dominant hardware solution in the hyperscale era?
In truth, it has, but not in the way that traditional server architectures have in the past.
According to MarketsandMarkets, the microserver segment is on pace to grow more than 53 percent per year for the remainder of the decade, jumping from today’s $1.9 billion to more than $16.3 billion. Top service providers, particularly in Asia, are increasingly turning to micro architectures in support of emerging mobile and cloud services while lowering infrastructure costs by eliminating power-hungry air-handling systems and other components.
This is obviously why companies like SuperMicro are churning out a steady stream of new models like the MicroBlade. The device is aimed at applications ranging from cloud- and web-hosting to video streaming and high-performance graphics and can support processors like the Atom C2000 up to the Xeon E5-2600. It also supports Ethernet connectivity from 1 to 10Gb and advanced VLP DDR4 16/32GB RDIMM. And perhaps most importantly for enterprises looking to convert legacy infrastructure to hyperscale, the MicroBlade fits in a standard 19-inch rack and can accommodate digital switching power supplies to help streamline support infrastructure as the number of blades increases.
With all of this going for the microserver market, though, why are some companies ditching the sector? AMD made the most notable move in recent months when it announced that it was exiting the business shortly after dropping a cool $334 million for SeaMicro. The word is that the company is more interested in developing SeaMicro’s fabric technology in support of its own emerging ARM processor line. But as of now, it doesn’t appear that AMD is selling the SeaMicro server business, but is simply shutting it down, and taking a $75 million hit to do so. This is odd behavior indeed for a business that stands to see 50 percent-plus growth over the next five years.
But perhaps the growth in the microserver market going forward is less than it seems. As a technology that is most conducive to the modular, scale-out architectures of web-facing giants like Google and Facebook, it just may be that the only winners on the server side will be the original device manufacturers (ODMs). As The Platform’s Timothy Prickett Morgan points out, even Facebook is still working out the kinks in the microserver piece of the Open Compute Project, and is working with both Intel and its preferred ODM, Quanta, to come up with an effective solution. Once that happens, expect the bulk of orders to go to ODMs, while the rest of the market will have to cut overhead to the bone in order to turn a profit on smaller-volume commodity businesses.
Microservers are indeed the right solution at the right time, but in the business world this is not always a recipe for success. There also has to be a way to produce the right solution quickly and efficiently and then sell it for a profit. And as enterprise infrastructure increasingly gravitates toward low-cost, low-power, widely distributed architectures, the old connections between manufacturer, distributor and customer will become increasingly frayed.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata, Carpathia and NetMagic.