I wrote a few weeks ago that “Microsoft taught the world how to succeed in PC and business software, but it might [be] teaching the world how to not succeed in cloud computing.” That’s a fate the company could avoid if it just delivered on a clear vision, and now it looks like Microsoft has decided on that vision: Treat cloud computing like it treated the PC business. (Oh, and it addressed some of those internal-cloud issues as well.)
The company earned its untold billions by developing an operating system that runs on third-party hardware. Through OEM deals, that strategy greatly increased Microsoft’s footprint and developer ecosystem and, thus, its importance. With the announcement of Windows Azure Appliances (WAP), Microsoft is once again looking to server makers to sell its software. But now it has added service providers to the mix too, and if Microsoft plans to make Windows Azure the Windows operating system of cloud computing, its “OEM” partners will be the key.
Just a week after the official announcement, the WAP marketplace is looking good. HP, Dell and Fujitsu are early partners, and all plan to implement WAP into their own data centers and eventually sell prepackaged WAP systems to their customers. Fujitsu has already committed to becoming a Windows Azure service provider too, first in Japan, then across the globe.
There’s reason to believe HP and Dell will follow in Fujitsu’s footsteps and sell Windows Azure as a service themselves. HP, in particular, already has announced plans to automate its data center in preparation to offer cloud services. Historically, most server makers have been content to develop the hardware platform but strike OEM deals for the operating system (along with other components). If application platforms are the operating systems of cloud computing, why not carry this practice over to the cloud? Large hardware vendors like Dell and HP can cloud-optimize their data centers while leaving platform development to Microsoft, VMware, Joyent and anybody else so inclined to sell their platforms to service providers.
Perhaps most telling about Microsoft’s chances is the report that Rackspace — which could be considered a Microsoft competitor in the cloud — is considering getting in on the WAP action. Like many hosting providers, Rackspace built its business on giving customers choice — colocation, managed hosting, cloud servers, Windows, Linux, etc. — so it isn’t such a stretch to throw a PaaS option into the mix. If customers want even more options, I’m sure VMware is open to talking about collaborating on a Java platform a la its partnerships with Salesforce.com and Google.
In competing for cloud computing dollars against Amazon Web Services, hosting providers like Rackspace have one big advantage: They haven’t made their cloud strategies on building every possible capability for every possible user by themselves; offering managed Windows Azure is just a natural evolution of offering Windows servers. AWS is building its own world of offerings (including, this week, Cluster Compute Instance), and it will build its own PaaS if it decides it needs such an offering. But Rackspace doesn’t have to undertake that development effort.
Of course, if this pans out at any notable scale, cloud computing could become business as usual for Microsoft. Is cloud computing a low-margin business? It is if you’re selling straight to developers. However, a large distribution channel for Windows Azure — a quality offering in its own right — will attract users. Hoping to capture these users’ dollars, more ISVs will build applications on top of Azure. Seeing a robust ecosystem, even more customers will follow suit. All of a sudden, Windows Azure is a must-have offering for service providers and Microsoft can reap the rewards just as it did with Windows operating systems. Right?