Microsoft Protects its Flanks to Save It's Ass

To say cloud computing is over-hyped is a little like saying ice ages are over-hyped. Cold, gradual, a change of only a few degrees but eventually cloud computing, like an ice age, will transform everything in its wake. One of the early signs of frost was made recently when Microsoft Chief Executive Steve Ballmer announced in Taipei that his company would spend $9.5 billion in research and development projects in 2010, a considerable portion of which will go to cloud computing technologies.

Mr. Ballmer was in Taipei to sign a memorandum of understanding with Taiwan's Ministry of Economic Affairs to set up a research center for cloud computing there, and to encourage local contract makers of electronics to develop devices related to cloud computing for Microsoft's Azure offering.

At stake is the high ground in the computer industry. Unlike enterprise computing, cloud computing is subject to both economies of scale and vertical integration. At the end of the day there will only be a few giant cloud providers with the potential to dominate the rest of the industry the way Wal-Mart dominates retail today. Who will be left standing - perhaps Google, Amazon, Salesforce.com, and Yahoo - is likely to occupy considerable management time and investor speculation in the decade to come. Unlike some of its competitors, most notably IBM and Oracle, Microsoft is at a disadvantage because it is not completely vertically integrated. Microsoft is dependent on other businesses to provide the hardware required to make its software run and the network equipment to hook them all together. Hence Mr. Ballmer’s visit to Taiwan.

Without control over the hardware and network layers of the stack, Microsoft is at risk of being frozen out from the engineering advancements that will make cloud computing more economical on a giant scale. To give a real world example of the new scale of some projects, the New York Times recently converted all of its issues from 1851 until 1922 from tiff to pdf format. Nothing earthshaking about that except the newspaper used cloud computing. In 24 hours it was able to process 4 terabytes of files (or about 4,100 trillion bytes) for a cost of $240.

Computing multiple projects on this scale goes far beyond the demands of the largest enterprise software customer and much of the engineering required to build, manage, and sustain it has yet to be developed. The early pioneers in this field have tended to eschew Microsoft products in favor of open source projects such as the Apache Software Foundation's Hadoop software (which interestingly was inspired by papers about Google's MapReduce and Google File System).

Not being competitive in cloud computing does not mean an end to Microsoft, but could threaten Microsoft's historically high margins. As a company, Microsoft is widely recognized for adopting a fast follower strategy toward technology. Unlike industry pioneers who bring new and transforming technologies to market, Microsoft is adept at improving existing technologies incrementally and is especially good at lowering their cost (although notoriously, sometimes at the price of reliability, which would be a fatal flaw in the cloud computing market). For Microsoft to spend something in the neighborhood of $10 billion dollars on the development of cloud computing over the next few years should alert investors to the real potential to create and lose fortunes in its pursuit.