Microsoft surprises by licensing ARM to create its own chips
electronista.com
electronista.com
As far as MS releasing a full .NET for other OSes? I can't ever see that happening. Why would they want their army of VB/C# "enterprise" developers being able to deploy without spending on Windows licenses?
Their support of Novell/Mono means people can still choose .NET for some projects; probably people that might not have used .NET otherwise. But at least the bulk of business apps developed will still need Windows (and probably SQL Server).
a) Without having a low cost deployment option the competiveness of microsoft is greatly reduced in the big data market. b) It might take Microsoft 3-4 years to work out the business model but having a portable .net framework with low / zero cost deployment would make apps written in .net viable in sectors where they dont have a chance currently - embedded , big data. c) The few startup people ive talked to have all worried about "theoretical portability", ie they might never use the portability but they would like the programs to work on different OS's. d) OSX actually benefits greatly from its unixy structural similarity to linux / bsd. Most developers I see here in NYC develop on osx and deploy on linux (linode, heroku etc) e) While Mono is a great project, having a blessed / direct version of .net would reduce the business risk in the mind of customers.
Microsoft was evaluating ARM for server usage, but do they need to build the SoC them self?
IIRC The last time they build some complex chips the result was far from perfect http://www.eetimes.com/electronics-news/4077187/The-truth-ab...
Xbox-style devices would sound more likely. The Xbox360 was successful in showing developers can make a bunch of low-performance cores perform adequately in games. I think a multi-core ARM coupled to a decent GPU would be quite a nice gaming platform.
I'm not sure if ARM is powerful enough to support a next generation console platform. The current Xbox 360 has 3 cores clocked at 3.2 GHz supporting two threads per core, while the most powerful ARM processor, ARM Cortex-A9 MPCore, has 4 cores that can each be clocked to "up to 1 GHz". Processors are a lot more complicated than just clock speed, of course, but I think it's readily apparent that ARM is less powerful than the current Xenon processor in the 360. Perhaps Microsoft could drop in several ARM processors or rely more heavily on the GPU.
I think it's more likely that this is a move to create a powerful in house ARM chip to use in their Zune HD and maybe for a to-be-announced mobile platform.
The only commonality is Windows itself. Windows NT on RISC processors has proved to be limiting in that almost no software was available to platforms other than Intel. This means the only functions a Windows on ARM server would be able to handle would be those of Windows itself and the software Microsoft choses to port.
It's kind of a copycat reflex. Apple is investing a lot in ARM, porting a big chunk of their OS to it, so Microsoft is convinced it must do the same, regardless of having a viable product.
So an ARM based server is something a lot of people is looking into (remember the SheevaPlug?) becase server are usually IO Bound.
Regarding a gaming device based on ARM, there are some but they are all portable and the most powerful available in the near future is the Nintendo 3DS (thanks to his GPU that is not ARM IP).
Here you can find a better analysis of the Microsoft-ARM deal: http://www.eetimes.com/electronics-news/4204864/Microsoft-AR...
Everyone else can play "me too" and copy Apple...