Q&A: Microsoft's View Beyond the Desktop

As chief research and strategy officer for Microsoft Corp., Craig Mundie is helping chart a new course for the software giant--one that takes it beyond the traditional PC world and into a universe of embedded, networked devices. Mundie's enthusiasm for this new space became evident as he keynoted a gala Windows Embedded CE 6.0 rollout last week in Redmond, Wash. Afterward, Mundie sat down with EE Times' Richard Goering to discuss moving beyond the desktop.

EE Times: The real-time operating system market is about $800 million today. Why is Microsoft interested in this relatively small market?

Craig Mundie: Because we see a future that has billions and billions of devices, all of which will eventually be connected. In the next few years, when almost every telephone in the world is going to be a smart telephone, when every device is connected, then the sophistication that will be required to deal with these things will go well beyond what's been available in the past. The second reason why this is important is that almost every product Microsoft has sold has a service component associated with it. The economics of the business goes beyond what you would get from embedded tool kits and operating systems.

EE Times: Who is Microsoft aiming at with this non-PC strategy?

Mundie: We have, for the most part, only sold products to the billion richest people on the planet. There's another 5 1/2 billion to go. Out of that, there's another 2 billion who are starting to have some disposable income. But the way they will get into the computing environment is probably different from the way you and I got into [it]. The first computers in their lives will be in phones, cars and televisions. Then they'll grow in terms of their demands and need other, more sophisticated computing capabilities. Everywhere you look, in order to fulfill peoples' needs, you can't do it all just with desktops and laptops.

EE Times: There's a large community of embedded developers out there. What impact will Microsoft have on them?

Mundie: I think we've already had a substantial impact, and I think the impact will grow dramatically in the next couple of years. The complexity of the systems that people are building today is wildly different from traditional embedded microprocessor systems. Each component is getting to be more complicated, and the ability to integrate them into a connected-device environment is getting to be more difficult, too. People are going to realize that taking a Linux kernel or something else and having their own people adapt it and extend it is going to be increasingly expensive. In the last 12 months, we sold as many units in our embedded-product line as we did in the prior nine years. It's an indicator that many embedded-systems devices are starting to enter the range of sophistication where our capabilities will be more fully appreciated.

EE Times: Can you potentially expand the embedded-developer community by giving people tools that make it easier to build these types of systems?

Mundie: We created Visual Basic and when we did that, the number of Basic programmers in the world grew by an order of magnitude, because it became easier to make applications. I think we're going to start bringing these visual programming techniques to the construction of sophisticated, highly concurrent embedded systems. I think that could bring a lot more people into the construction of embedded systems.

EE Times: We're hearing a lot about multicore CPUs and systems, and the difficulty of programming them. Is multicore support in your road map?

Mundie: Absolutely. I've been responsible for coordinating a lot of Microsoft's research and incubation efforts in anticipation of the arrival of high-core-count CPUs and heterogeneous system-on-chip architectures. I think the company is extremely well-prepared for the emergence of these technologies, and we will bring forward what we think are optimal solutions, both in the embedded space as well as traditional desktop and server product families.

EE Times: Any embedded multicore solutions yet?

Mundie: No, in the embedded space they haven't emerged as a big factor yet. [But] the embedded community will not be immune from this challenge.

EE Times: Why are you opening the Windows Embedded CE 6.0 kernel?

Mundie: More than a decade ago, we started making limited source code licenses of Windows available to research professors. Then, as the open-source com- munity grew, we realized in 2001 that we had a whole bunch of ad hoc ways where different parts of the community had been given access to different parts of our source code. We came up with a set of licenses tailored to the different usages that people needed for source code. With this [6.0] release, virtually all the kernel is available. That is largely to support the community's requirement for debugging.

EE Times: What types of embedded systems will Microsoft seek to facilitate?

Mundie: More and more, we think many systems will have a part of their value created by a connection to Web-scale services. . . . For that to happen, they need to speak a common set of protocols to get at those Web services. Another thing that's going to come to the embedded space is changes in modality of the user interface. There will be speech and touch and vision and other kinds of mechanisms where the embedded system will deal with people. All of those will call for more horsepower in embedded systems, as well as a much higher level of tools and libraries than were historically part of this space. These are all areas where the company has invested for a long time.

Craig Mundie

Born:
1949, Cleveland

Education:
Bachelor's degree in electrical engineering (1971) and master's degree in information theory and computer science (1972) from Georgia Tech

Current position:
Microsoft Corp.'s chief research and strategy officer since June, responsible for long-term strategy and technology vision

Career:
Joined Microsoft in 1992 to create and run the Consumer Platforms Division, which was responsible for developing non-PC platform and service offerings such as Windows CE; software for handheld, pocket and auto PCs; and early telephony products. Named senior vice president of consumer strategies in 1998 and chief technology and strategy officer in 2000

Co-founded Alliant Computer Systems in 1982, later served as CEO.

Started career at Systems Equipment Corp./Data General; developed Secos operating system for DG Nova