Microsoft Expands On “Software Plus Service” Concept

I sort of “live blogged” the Microsoft (MSFT) Financial Analyst Day. I say, “sort of” because I did not actually post it up in real time, and I was in and out of the room with phone calls. And, oh by the way, I was not even there (but attended over the web).

In the afternoon session, we heard more about the Software Plus Service strategy than we had heard in the spring. More important, we heard less about advertising and devices being of equal importance with software and service than we had heard in the morning sessions from Steve Ballmer (comments on the morning session are here).

In the morning Ballmer had talked about the four (software, service, advertising, devices) as if they were the pillars of Corinth (I know, there are only three remaining). But by the time of the Q and A at the end of the day, advertising had returned to being a way to monetize technology, as it is with Google (GOOG), and I didn’t hear a mention of a Microsoft telephone or Microsoft TV.

Forgetting Ballmer’s morning foray complicating the discussion, the issue is that even a “software PLUS service,” as opposed to a software as a service (SaaS) approach, potentially extends Microsoft beyond what it can afford to do. The good news is that Ray Ozzie and Craig Mundie, Microsoft chief architect and technology strategist respectively, talked about “software plus service” as if it was SaaS.

Both alluded to previous-life experiences that firmed up their understanding of what best belongs “in the cloud” and what belongs “on the edge.” Ozzie’s story, especially related to Notes, is well known. Mundie developed some breakthrough data management software back in the day and was one of the founders of the company that brought the successful Alliant parallel processor to market in the 1980s. Probably less well known is that Mundie (like Ozzie, a Data General alum), is the guy that strategized the Data General software that never had a chance in the market because it ran on the system called Fountainhead [FHP] that was never marketed, as documented in the 1980-Pulitzer-Prize winning book, The Soul of a New Machine.

Ozzie said he thinks of every Microsoft software offering as a “socket for a service.” Ozzie is theorizing (and helping to enable) a services platform approach for Microsoft to provide speed, scale and monetization (and cost savings for the IT budget) to both Microsoft itself and Microsoft’s enterprise customers. The platform he describes is the utility computing concept as MIT, GE (GE) and Bell Labs first conceived it in the late 1960s. See his speech for a full breakout of the framework.

Mundie said the 20-year personal computing model is evolving. He couldn’t say it’s dead of course so he pictured it as a pendulum swinging back and forth between personal and centralized computing every 20 years (now being the time it swings away from personal). Doing a great Carl Sagan, Mundie began a software discussion by explaining the implications of “watts generated per centimeter” of chip real estate. It’s the basic thing that is holding back Moore’s law as Gates had talked about in the morning (as opposed to physics). It’s all about heat. In order to compute more things in the current hardware-design paradigm, you need higher-frequency computing devices. Moore’s law will allow it. But at some point you would need asbestos gloves to work the mouse and asbestos pants to hold it on your lap.

The solution Mundie says is specialized chips that will show up in all kinds of devices scattered all over the framework described by Ozzie, in the cloud and on the edge. This means that the free lunch of higher and higher performance is over for software developers. Programmers that are now used to a one-size-fits-all programming approach will have to “write to” specialized chips. The software of tomorrow will have to take advantage of a loosely coupled (Software PLUS Service), asynchronous environment based on composite layering, decentralized compute power, and resiliency achieved via redundancy.

Such a brave new world will balance the personal computing vs. centralized computing pendulum, stopping it from swinging. Microsoft wants to sit right in that middle and is well positioned to do so (“in the pole position,” according to one analyst present). But it has to decide whether it wants to be the utility, or the guy that makes the dynamos. “Software plus service” is a costly conflicting corporate strategy that may become a self-fulfilling prophecy, a step back to the 1990s. Hell, a step back to the GE of 1910. If Microsoft tries to straddle that fence (be both a service provider and a technology supplier), its investors will be the ones hit by the pendulum.