Is it time to rethink Windows?

[Disclosure: Many of the firms mentioned are clients of the author.]

It’s easy to forget that Windows was initially a shell for DOS before it went through a massive change in 1995 (after Apple made GUIs the next big thing). Microsoft dominated the market that resulted, of course. But this was a market defined by desktop hardware, where servers sat in the background and the idea of a terminal was mostly verboten. {Before the rise of the PC we had terminals tied largely to IBM mainframes that were far easier to maintain, arguably more secure because they couldn’t run viruses, and far more appliance-like.)

As we move to a cloud-driven present where offerings like Azure define the space, shouldn’t we clean-slate a new client much as we did with DOS, Windows and iOS? Doesn’t it make sense to strive for a result optimized for the cloud world of tomorrow rather than the PC world of yesterday?

Clearly the Windows Virtual Desktop showcases the need for this change. But much like we didn’t take the operating system off IBM’s mainframe and port it to PCs, I don’t think we should take the operating system off of PCs and shoot it up to something that looks a lot more like a mainframe today.

Why we might want to rethink windows

Windows goes back to its DOS roots, and every generation has built off the generation before it. This layered complexity allows applications to migrate and provides users with a level of comfort that they wouldn’t get from a changed product. But it also takes constructs that were for different generations of technology – generations that are now obsolete – and passes them up to current versions where they undoubtedly slow performance, increase security exposures and increase the costs related to updating and maintaining the offering.

A company like Microsoft can, and does, handle this overhead, but should result in unnecessary costs and problems. These problems would be avoided if the firm took a tabula rosa approach from time to time to get rid of all the baggage and start anew.

If you think of the things like drivers that have historically defined the support problems with Windows in a cloud environment, they should go away because you virtualize the hardware, which means you don’t have to adapt the hardware manufacturers anymore. Most of the driver problems shouldn’t exist. Rather than having something like VMware with Windows layered on top, why not just have a blended hypervisor that was also a complete OS and avoid the layered complexity?

And I’m not just talking one OS either because you don’t need the same OS in the Cloud that you have on the device. For the device, you need something more like a Thin Client OS so that the complexity and maintenance headaches of a heavy client go away.

Timing and IBM’s quantum example

We don’t have pervasive high-speed wireless connections yet – and we won’t until 5G reaches critical mass. But we’re a couple of years out from that, and we’re already exploring new hardware like the joint Qualcomm-Microsoft Always Connected PC efforts.

But if you look at what IBM is doing with quantum computing – technology that’s at least five years out – they’ve already engaged developers so that when the technology is ready the ecosystem will be ready to take it. This trend means that Microsoft, if it wants to make the pivot timely, should be building for it now, and there is some risk if they don’t do it in a timely fashion.

The China threat

That risk is China. Were I one of that country’s analysts, I’d be using Apple as an example of how to take the market away. Recall that Apple entered the smartphone segment that was resisting a move to make consumer-focused smartphones (efforts that preceded Apple’s in both Palm and Microsoft were killed by their respective firms) and took it over.

Before Apple the big players in smartphones were Nokia, Palm, Research In Motion (Blackberry) and Microsoft. After Apple’s success Nokia was mostly gone, Palm was gone, Microsoft was out of the phone business and Blackberry changed their business model away from smartphones. If Apple had been a Chinese rather than US company, the smartphone market would have pivoted to them, and Google, given they’d have been fighting China, would have likely not even tried to create a competitor to avoid premature battle with China.

We are at another inflection point, and while China is currently blocked, they are finding ways around that block with Huawei. Given they have a ton of leverage (they own most of the US debt), they still have a powerful path to success. If China were able to replicate what Apple did with the smartphone with a cloud service and overcome the politics, that one move could move the center of the tech market from the US to China. It’s critically important that a US company get this right first if the US wants to remain the world’s tech leader.

No coming back

The rise and success of platforms like Azure and the IBM cloud, not to mention Virtual Windows, indicate the need for two clean-slate operating systems. One for the cloud that builds on top of a hypervisor and one on the client that completes the solution. This approach would seem an obvious path to repeat what Apple did on smartphones and Apple, Google or any other OS competitor could replicate it. But the real risk as China advances is that they may figure this out first and create a foundation for that stated plan to move the tech segment from the West to the East. Once moved, it will be nearly impossible for it to come back.

IBM didn’t really learn from Microsoft and even though the IBM PC largely created the PC market, IBM had to exit it as a result. Maybe Microsoft should learn from IBM this time, so they don’t repeat what would be an ironic recurrence of that past.