Directions

Microsoft Faces Reality About the Future of Computing

Will the desktop computing model last into the next decade? Microsoft thinks so, yet it's begun to prepare for a future where desktops reign no longer.

• By Paul DeGroot
• September 01, 2007

In the face of steady defection of both consumers and businesses to network-based applications, Microsoft has insisted that the desktop is still the core of personal computing. The main technical arguments behind that included the vast supply of computing cycles on the desktop, local storage, offline applications and the security of desktop data.

None of those arguments is particularly compelling. The biggest bottleneck for most people is not CPU or disk capacity, but their network connection, and even a slow broadband connection can deliver adequate performance for 95 percent of Internet use on any computer built in the last five years.

Local storage can actually be a drawback in a mobile and multi-device world, where it forces users to synchronize locally stored data among various storage locations or devices.

As Wi-Fi and better technologies, such as WiMAX, become ubiquitous, the offline requirement lessens. Microsoft likes to cite air travel as an example of wasted time because people can't get online during flights, but in reality, many travelers fill the gap with other productive pursuits such as reading. Meanwhile, of course, battery life remains a major constraint. Furthermore, this gap is likely to disappear within the next five years as airlines offer in-flight connectivity (assuming that they don't also eliminate the gap between seats).

Many future applications (based on technologies such as Adobe's AIR, JavaFX and Microsoft's Silverlight) will work transparently offline and online.

Finally, the desktop isn't necessarily safer than an online repository for storing important data. Keystroke loggers can get you either way, and locally stored data is actually more vulnerable to the most common threats, such as spyware or viruses that enable remote control.

The biggest threat to Microsoft hasn't been Google, YouTube, eBay or other Internet heavy hitters. Instead, it's been the company's unwillingness to depart from a hugely profitable computing model that no longer has a long-term (i.e., beyond 10 years) future. That's understandable from a business point of view, but that view was myopic and Microsoft's management needed to get its head out of the sand.

We got that at the WPC. Ballmer outlined, as clearly as Microsoft has ever done, how Microsoft would make network-based computing as good as or better than today's local computing architecture. He discussed a new computational model in which "geo-scale" centers (ironically, the mainframe redux) do the heavy lifting.

That said, Microsoft's vision is far from perfect. For one thing, Ballmer's technical architecture for "cloud computing" looks suspiciously like the architecture for local computing. Microsoft's imagination may be hobbled by what it already knows.

In addition, a business model was nowhere in sight. Network-based computing eliminates Microsoft's single biggest business advantage, the OEM relationships that let the company populate the world's computers with its software faster than any competitor can. Much of that advantage disappears when apps and data aren't resident or preinstalled. This threatens the huge monopoly profits that have hidden so many weak product lines and bad business decisions. That Microsoft can compete, I am confident. That it can compete profitably, I am not.

Finally, beyond some cheerleading about partners, the channel model was invisible in Ballmer's remarks.

It's time for partners to do what Ballmer has done. Take a hard look at the future and figure out where you fit. I'd offer two possibilities:

First, if you have ideas for how Microsoft can cross the chasm ahead, now is the time to articulate them. Microsoft may be more willing than ever to listen.

Second, don't count on Microsoft to take you across. Your best opportunities may lie elsewhere.