CNET también está disponible en español.

Ir a español

Don't show this again


Microsoft's Manhattan Project

Microsoft chief software architect Ray Ozzie is leading a risky but unavoidable Manhattan Project effort to build a planetary scale cloud computing platform.

This week Microsoft gave evidence that it will continue to be a major force for at least the next decade. The company outlined its products and strategies that more fully embrace the "cloud," such as the Azure set of cloud services; Web-based, lighter-weight versions of Microsoft Office applications; and the latest iteration of the Live Mesh middleware. Google may have won the search war, but Microsoft isn't about to cede the global cloud to the search engine giant.

Ray Ozzie explains Azure to CNET News correspondent Ina Fried.

As in past epochs in its 33-year history, Microsoft ties its success to the developer community, having an army of loyal, or at least well or modestly compensated, software warriors. The Microsoft mantra is: "Build a platform and an ecosystem of developers, partners, fans, and people willing to spend their money will follow." A compelling platform and the potential to reach a large audience of buyers, which Microsoft can deliver, attract the developers, who build the applications and services that attract consumers and business users.

Microsoft also now understands that its platform must span every kind of device--PC, notebook, smartphone, car, home, etc.--and offline scenarios. Microsoft missed the Web search revolution, but it's not going to miss out on the much bigger revolution--the move to the cloud over the next two decades.

Google is building a competing ecosystem from the ground up with similar characteristics and a desire to attract millions of developers. Amazon is pushing its elastic computer cloud, and Rackspace, EMC, IBM, and many other companies are trying to get a piece of the action. Most the cloud companies are focused on hosting services, but the biggest piece will be platforms-as-a-service with developers creating and running their applications for on a cloud operating system.

An early example of this trend is's proprietary platform. Sun Microsystems, the company that coined the phrase "The network is the computer," has all the pieces to construct a planetary cloud but seems to be missing from the discussion. As my friend Steve Gillmor notes, Sun is on the ropes.

Openness is a major issue as the global cloud materializes. Businesses don't want to be locked into a particular cloud, and also want various clouds and services to interoperate via standards. Speaking at the Professional Developers Conference last week, Microsoft's chief software architect Ray Ozzie said that the foundation level in the operating system cloud would run in Microsoft's data center, but SQL services, .NET, and Live services can be mixed and matched by developers inside and outside of the company's datacenters. The Azure cloud is also cross-platform, but the various clouds will extract a toll and by nature it won't be dead simple to move applications using foundation services from one cloud to another.

Microsoft's cloud computing efforts have gotten off to a slow start compared with competitors, and it's on the scale of a Manhattan Project for Windows. Azure is in pre-beta and who knows how it will turn out or whether consumers and companies will adopt it with enough volume to keep Microsoft's business model and market share intact. But there is no turning back and Microsoft has finally legitimized Office in the cloud.

Ray Ozzie has a track record of slowly but surely getting things done and Microsoft is famously persistent and cash rich. But building a platform, or Internet operating system, at planetary scale supporting billions of users and trillions of transactions per day, and having fleet Google as a primary competitor will be a major test of Microsoft's brain trust and resolve. Don't be surprised to find a recharged Bill Gates parachuting into the fray as Azure evolves and the cloud war for developers escalates.

See also:

Scoble: Never underestimate Microsoft's ability to turn a corner

Wilcox: How Can You Be So Sure about Azure?