My blood boils every time I hear some Microsoft executive crowing about its so-called 'software-plus-services' model. Microsoft's FUD around S+S is deliberately muddling up two very different things. I fumed when I read this comment in an interview last week by Chris Capossella, senior VP of the group that oversees Office, in which he singles out Google's use of Gears technology to enable offline use of its cloud-based applications:
"I thought the whole point [of the software-as-a-service model] was that I didn't have to download anything," Capossella said. "These guys are totally adopting the software-plus-services approach, but they just aren't using the term. And no one's calling them on it."
This is so lame as a line of attack, and yet it's a favorite at Microsoft, as if Google's decision to take advantage of the compute power that's there on the client justifies in one fell swoop not only Microsoft's entire legacy stack of desktop-bound applications and operating software but also the whole gamut of its extended server family. Extending cloud-based services so that they'll run locally in a few limited use cases is in no way equivalent to Microsoft's policy of encouraging its customers to keep buying and upgrading their installed base of server and desktop software in return for assurances that the vendor has a strategy of offering the "choice" of cloud-based equivalents.
My opposition to the 'software-plus-services' mantra is that it puts the cart before the horse. If you're going to do cloud computing right, you have to start with services. I wouldn't be debating this with Microsoft if the company were arguing for 'services-plus-software' — or even 'software-powered services', which is a phrase I was using to describe SaaS back in 2004. What's important is to shift from a software product mentality to a software-enabled services mindset.
For all the bluster of its senior VPs, it turns out that the next generation of Microsoft's flagship server products is being designed in exactly the way I've been arguing for. In a little noticed interview with CNET's Ina Fried last month, corporate VP Rajesh Jha revealed that Exchange 2010 (formerly known as Exchange 14) — along with many of its Office 2010 stablemates — is being proven as a service first before later on being turned into a server product:
"The last time around, though, Microsoft built the server software first and then delivered the service. In developing Exchange 14 — and indeed many components of the next Office — Microsoft has flipped the switch and is instead developing the service first and doing the server work second."
Behind the scenes, then, Microsoft has begun the technology re-architecture that will transform it into a cloud services company. I'm sure that, once it has completed the transformation, its users will still be able to use some of that functionality when disconnected (or semi-connected, as Gmail's 'flaky mode' now enables), and it will still sell server software to customers who choose to act as providers of cloud services in their own right. But the central tenet of the 'software-plus-services' mantra — that the average enterprise wants to be able to move its application infrastructure into the cloud and back out again at will — will have long since been disproven. For as Microsoft's Online Services division is discovering, the traffic is all one-way. Once customers have moved to the cloud, they never want to leave.