Intel can really hit back by releasing very fast chips again, so fast that emulation and virtualization becomes completely painless, rendering Microsoft's OS planned obsolescence completely moot because then anyone can run any version they choose. They could really hurt them by getting into their own OS business, even if it is merely a super virtualizing hypervisor that sits between the hardware and any Windows version. This would be Microsoft's worst nightmare.
This is interesting -- can you elaborate on the part about a virtualizing hypervisor being (potentially) Microsoft's worst nightmare? I can see how Intel releasing its own OS could be a problem for MSFT, but am not so clear on how the hypervisor thing would affect MSFT. Why would they care -- couldn't they argue that unless you're running the latest version of Windows, you're exposed to Internet nasties and also missing out on great new features?
Intel very well could write a complete x86 operating system, in fact I would say it is likely they already have for internal use because I simply cannot imagine them needing Microsoft to supply one for their development systems holding schematics, roadmaps and other top secret stuff. The danger for them writing an OS for consumers has at least two big problems ...
(1) The final user interface design may drift too far away from classic Windows ( like Metro! ) and turn off the market and it would be dead in the water after a substantial development and testing expense. What would be the point of a nice OS that is GUI crippled to end-users? They have shown they won't touch it.
(2) The legal war with Microsoft over the API. Even though Intel owns the intellectual property of the processor instruction sets, Microsoft no doubt holds patents on the API implemntations. This is mega-complicated by some of them like Win16 and Win32 being way old but probably adjusted along the way so that they somehow are kept out of the public domain. My gut feeling is that Intel would win on the merits but lose financially from the expense of the legal bloodbath. It is a fight that needs to be fought though.
So if the goal is for the end-users to be able to continue using their own Windows versions on existing and forthcoming hardware ( and flip the bird at Microsoft because they are using NuWindows as a trap ) then it might be a better idea to forgo a cleanroom clone of Windows ( because they just might get the thing working great but screw up the GUI ) and just develop the low level software that facilitates the already existing Windows operating systems.
VM's started out functioning at much lower levels than we see today with most running essentially as just another piece of client software ( although I see that Wikipedia says that Hyper-V is considered low-level ). It will make more sense if I just copy that section here ...
In their 1974 article "Formal Requirements for Virtualizable Third Generation Architectures" Gerald J. Popek and Robert P. Goldberg classified two types of hypervisor:
Type 1 (or native, bare metal) hypervisors run directly on the host's hardware to control the hardware and to manage guest operating systems. A guest operating-system thus runs on another level above the hypervisor.
This model represents the classic implementation of virtual-machine architectures; IBM developed the original hypervisors as bare-metal tools in the 1960s: the test tool, SIMMON, and CP/CMS. CP/CMS was the ancestor of IBM's z/VM. Modern equivalents include Oracle VM Server for SPARC, Oracle VM Server for x86, the Citrix XenServer and VMware ESX/ESXi.
Type 2 (or hosted) hypervisors run within a conventional operating-system environment. With the hypervisor layer as a distinct second software level, guest operating-systems run at the third level above the hardware. VMware Workstation and VirtualBox exemplify Type 2 hypervisors.
The classification of specific hypervisor implementations as Type 1 or Type 2 is not always clear cut. For example:
- Kernel-based Virtual Machine (KVM) is implemented as a kernel module for Linux 2.6.20 which, when loaded, allows the Linux kernel to operate as a bare-metal (i.e., Type 1) hypervisor. However, as Linux is an operating system in its own right, one can argue that KVM is a Type 2 hypervisor.
- Microsoft Hyper-V (released in June 2008) has also been misidentified as a Type 2 hypervisor. Both the free stand-alone version and the version that is part of the commercial Windows Server 2008 product use a virtualized Windows Server 2008 parent partition to manage the Type 1 Hyper-V hypervisor. In both cases the Hyper-V hypervisor loads prior to the management operating-system, and any virtual environments created run directly on the hypervisor, not via the management operating-system.
Attempts have been made to introduce the term Type 0 (Zero) Hypervisor to differentiate specific hypervisor implementations. However, no consensus as to the validity of this term has been reached.[/b]
Now, Intel is closer to the CPU instruction set and the microcode than anyone on Earth, and in fact has even added VT-x ( see this for details ) to some of their CPU's, unfortunate key word here being some. Physically all the pieces are in place for a perfect bare metal solution if Intel would stop playing games by only adding features to certain chips to segment their products.
That perfect solution ( for consumers tired of chasing Microsoft but want Windows ) is a low-level, lean and mean hypervisor that on one side talks right to the chip microcode and on the other side presents the normal instruction set to the operating system. This could be a 2nd, higher layer of microcode and therefore might be just added to the CPU itself or a secondary on-package chip. The existing operating systems would still see the CPU and retrieve the normal CPUID to determine which HAL to install and you have business as usual after that.
Naturally this depends upon Intel's benevolence and sympathy for the plight of the x86 universe which Microsoft is presently targeting for destruction ( they are not trying to kill the physical architecture, but are planning on preventing user mode software from ever talking to it again ). This is why I say it would be Microsoft's worst nightmare because it removes them from the new operating system business altogether. Users could install whatever version of Windows they have lying around and hardware makers would only need to write device drivers for new hardware using any existing DDK targeting any version of Windows they want.
Of course there is much wishful thinking here, and the point is that these thoughts never crossed our minds until now. Microsoft is getting out of the neutral operating business and moving to a curated gatekeeping system. The very best solution would be for Microsoft to liberate all the x86 source code and related IP and continue with their walled-garden Metro madness. The 2nd best would be for the government to send in SWAT teams and just take it ( which would be an ironic taste of what some people have actually had happen to them thanks to IP and patent laws pushed by Microsoft and others ).