ARM Servers. A challenge to the WinRT idealogy?
Back when I first read that Microsoft decided not to enable WinRT to run 3rd party desktop apps their justification was that these apps don't have an foundation in place that stops them from consuming battery, and that ARM processors simply weren't powerful enough to make the experience enjoyable.
Since then, many Windows desktop apps have been recompiled for ARM and run "jail-broken" Surface and other WinRT devices. And guess what, they run fine. I'm not saying that there won't be apps that don't run great. But I own an x86 tablet with an Atom processor in it. By all accounts it is on par with or in some cases even weaker than ARM tablets. But Microsoft has no beef with allowing me to run that software on my tablet.
Amusingly, while I could probably, to Microsoft's point, find applications that run poorly or rape my battery life. The reality is, I rarely use desktop apps anyway. The touchscreen was not designed for them. So, aside from the brief foray into the desktop environment when I have no other option, the only time I'm really using that environment is when my tablet is docked. And that eliminates one half of the problem. The battery life. But, to be honest, the battery life, the few times I've used a desktop application extensively, doesn't even seem to have been impacted. At least not noticeably.
I had argued when I heard about this other component, the performance argument, that it was short-sighted. At the time Tegra 3 was on the horizon and Tegra 4 already had specs. And Tegra 4 already looked like a viable competitor, performance wise, for anything Atom branded. And I think it delivers on that. So, Windows 8 has been out barely a year and we are already seeing high end ARM chips that beat out the lowest end x86 chips. And lets not pretend that Microsoft didn't know these were on the horizon, they work directly with their OEM's and get advance prototypes of hardware to test and certify on.
Ok, so Microsoft's goals were a bit misguided. Or they lied. Anyway, for the consumer market, I don't think they would bother going back on that decision unless they felt that they absolutely needed too. But, AMD just announced a powerhouse ARM based, Opteron branded server chip. And ARM just announced standardized architecture designs for ARM based servers.
Suddenly, this could be bad news for Microsoft's decision. If ARM based servers starts hitting the mainstream, Microsoft needs an OS which is enterprise grade that runs on ARM and provides ALL of the functionality people have come to expect if they hope to maintain their control over the enterprise segment.
There are already server grade Linux OS's that are much more complete than Microsoft's WindowsRT based offerings. And the UI's get more IT department friendly every year.
Don't get me wrong. This isn't sudden death. Enterprises tend to be very slow moving indeed. As a result I doubt that there would ever be en masse departure from Microsoft's current x86 based offerings. But, over time, new companies will adopt the platform with the image of being better on energy consumption and will simply hire around ability to work with the OS's that run on it. And, much later on, existing enterprises will eventually need to replace hardware and upgrade/change OS's and at that time, even they may make the move.
The enterprise segment is a tricky one. It moves much slower than virtually every other tech segment, and it can take years for a mistake to play out or a good move to return on investment. That pace also makes the relevance of any one single move lessened over time. There isn't any telling what might happens in the years before these things come to full fruition to change them. But, I feel like, if ARM starts getting properly into the server market and Microsoft doesn't have a viable OS to answer with, that we'll see market share erode drastically.
Since then, many Windows desktop apps have been recompiled for ARM and run "jail-broken" Surface and other WinRT devices. And guess what, they run fine. I'm not saying that there won't be apps that don't run great. But I own an x86 tablet with an Atom processor in it. By all accounts it is on par with or in some cases even weaker than ARM tablets. But Microsoft has no beef with allowing me to run that software on my tablet.
Amusingly, while I could probably, to Microsoft's point, find applications that run poorly or rape my battery life. The reality is, I rarely use desktop apps anyway. The touchscreen was not designed for them. So, aside from the brief foray into the desktop environment when I have no other option, the only time I'm really using that environment is when my tablet is docked. And that eliminates one half of the problem. The battery life. But, to be honest, the battery life, the few times I've used a desktop application extensively, doesn't even seem to have been impacted. At least not noticeably.
I had argued when I heard about this other component, the performance argument, that it was short-sighted. At the time Tegra 3 was on the horizon and Tegra 4 already had specs. And Tegra 4 already looked like a viable competitor, performance wise, for anything Atom branded. And I think it delivers on that. So, Windows 8 has been out barely a year and we are already seeing high end ARM chips that beat out the lowest end x86 chips. And lets not pretend that Microsoft didn't know these were on the horizon, they work directly with their OEM's and get advance prototypes of hardware to test and certify on.
Ok, so Microsoft's goals were a bit misguided. Or they lied. Anyway, for the consumer market, I don't think they would bother going back on that decision unless they felt that they absolutely needed too. But, AMD just announced a powerhouse ARM based, Opteron branded server chip. And ARM just announced standardized architecture designs for ARM based servers.
Suddenly, this could be bad news for Microsoft's decision. If ARM based servers starts hitting the mainstream, Microsoft needs an OS which is enterprise grade that runs on ARM and provides ALL of the functionality people have come to expect if they hope to maintain their control over the enterprise segment.
There are already server grade Linux OS's that are much more complete than Microsoft's WindowsRT based offerings. And the UI's get more IT department friendly every year.
Don't get me wrong. This isn't sudden death. Enterprises tend to be very slow moving indeed. As a result I doubt that there would ever be en masse departure from Microsoft's current x86 based offerings. But, over time, new companies will adopt the platform with the image of being better on energy consumption and will simply hire around ability to work with the OS's that run on it. And, much later on, existing enterprises will eventually need to replace hardware and upgrade/change OS's and at that time, even they may make the move.
The enterprise segment is a tricky one. It moves much slower than virtually every other tech segment, and it can take years for a mistake to play out or a good move to return on investment. That pace also makes the relevance of any one single move lessened over time. There isn't any telling what might happens in the years before these things come to full fruition to change them. But, I feel like, if ARM starts getting properly into the server market and Microsoft doesn't have a viable OS to answer with, that we'll see market share erode drastically.
Comments
Post a Comment