Microsoft gave more details Tuesday on its plans to launch Windows-based “hypervisor” software for running multiple operating systems.
Microsoft ‘hypervisor’ plan takes shape
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Maybe we’ll have the oportunity to run linux and windows with a MS boot loader ! hihi
Xen has certainly caught a lot of people knapping, including it seems, Microsoft. The amusing part is that it doesn’t yet run with Windows, and from a strategic point of view, I hope that continues. Of course, it’s then just like Microsoft to say that the market is in its early stages.
Maybe we’ll have the oportunity to run linux and windows with a MS boot loader ! hihi
Can’t find NTLDR
Yer, that sounds like something I really want to see a lot more of.
Xen has certainly caught a lot of people knapping
That should be napping of course, but come to think of it, maybe Microsoft should go back to basics and start making flint tools.
I suspect that MS Hypervisor will actually be better than VmWare or Xen just because they know how to make it work with Windows and since BSD/Linux/Solaris are now open source, it should be easy to provide excellant support for *ix. The only wildcard is MAC-OSX – but they’ll cross that bridge when the OSX gets there. At the same time I also think that their software will suck for a couple of versions (like every thing that MS makes).
The article fails to mention a few things:
1. Xen currently runs on x86 hardware, so you don’t have to wait for and upgrade to expensive hypervisor hardware. Xen is available as source and also packaged with (at least) SuSE (out now) and Fedora Core 4 (due Monday). I think there’s also good support for Debian and NetBSD.
2. Xen supported Windows previously, but Xen had to cut Windows off because Microsoft wouldn’t budge on licensing issues. (Compare this to Microsoft’s recent announcements: they are great on compatability and they are interested in working with open source software. And compare this to Microsoft’s monopoly problems in the EU.)
3. With hypervisor hardware, I think Xen will support Windows.
4. Microsoft will probably be charging for their product, while Xen is free of charge.
5. Xen runs your choice of operating system as Domain 0 (the small, master system). That means, if you want, you can choose operating sytem X because you like its traffic-shaping firewall.
I (and the author) should have looked in the Xen FAQ. So the answer is that Xen will support Windows as soon as Microsoft’s product does. (Both are waiting for the new hardware.)
http://wiki.xensource.com/xenwiki/XenFaq#head-fcb85a149da66907086cc…
<pre>1.4. Does Xen support Microsoft Windows?
Unfortunately we do not currently support Windows; the paravirtualized approach we use to get such high performance has not been usable directly for Windows to date. However recently announced hardware support from Intel and AMD will allow us to transparently support Windows XP & 2003 Server in the near future. We are working on this and intend to have support available by the time the new processors are available. </pre>
Sorry, one more post.
(Xen is really cool!)
Xen lets you migrate a running system from one physical computer to another in real time! This makes it easier to achieve super long run-times for critical systems.
Xen 3.0 is due out next month will great new features such as native 64-bit support. See http://www.cl.cam.ac.uk/Research/SRG/netos/xen/roadmap.html
ironically, microsoft played a part in the early development of Xen, and they opened their code to the xen developers – and so they did have a version of windows ported to Xen. however microsoft decided they wanted no part of that anymore … just like the internet was a passing fad …
Does Xen support OSX?
The Xen home page and FAQ show Xen does not support Mac OS X. Xen requires an x86 processor.
However, a few things:
1. FAQ states PPC support is planned in due course.
2. FreeBSD support is in the works, so Mac shouldn’t be too hard.
3. Now Mac is coming to x86.
This annoucnement is ironic considering in the now infamous “Halloween Documents”, the M$ rep bashed F/OSS as being “backwards looking”.
I’m not so sure the MS Hypervisor will require hardware virtualisation extensions. It’d certainly make some kind of engineering sense if it did but I had rather assumed they’d leverage the VirtualPC binary rewriting tech they already have to provide virtualisation on vanilla x86 hardware.
The article doesn’t make it clear and I suspect we’ll have to wait a while to get a definitive answer: I suspect even MS may not know for sure at this stage!
The WinHEC slides (which as usual, CNet does not think relevant) say that Windows Hypervisor requires VT/Pacifica.
Thanks very much! I missed that when I originally skimmed through.
It makes sense: by 2007, these extensions will be pretty widely available anyhow, so there’s no sense incurring the pain of dynamic binary rewriting.
A port of OS X to PPC Xen was planned but nobody’s working on it right now. Likewise, a port of x86 OS X to Xen is perfectly possible, just not done.
With Vanderpool / Pacifica support, Xen will do full virtualisation, so eventually all PC OSes should Just Work (TM).
Microsoft is no stranger to virtualization techniques. MS OS/2 1.x running on 286s used the virtualization techniques of the 286 chip, not not very efficiently – the 286 was not an efficient chip as regards virtualization.
OS/2 2.x used the 386 virtualization much more efficiently, but then MS and IBM split up. And the 386 virtualization wasn’t very extensive anyway – merely a set of virtual 8086s.
MS WinNT had a stab at putting a virtual chip at its foundation – the HAL. Most of the upper levels of ntoskrnl communicate to the HAL and through the HAL to the hardware, so it shouldn’t be very difficult for Microsoft to achieve.
But it’s taken the pressure from Xen’s success for them to try this now? Buncha slackers.
First of all i would like to say to all the OSS zealots that they if they feel that XEN will be the winner or its way cooler than other two technologies then they shouldn’t need to come and brag about it..correct?
Secondly, XEN the last time i checked was unstable as a 30 year old car. No no don’t mess with it…it will crash..
VMWare and Virtual PC/Server both are rock solid. Which means both EMC and MS are a player in the market.
Now lets where everyone stands:
MS Hypervisor and XEN are almost seems like the same technology…run the hypervisor on bare hardware, run a domain or partition with device drivers for hardware and service other guest os from that domain.
VMWare on the other hand has ESX which is hypervisor + device drivers.
VMWare has the limitation that as new hardware comes, it will have to add new device drivers to its hypervisor, making hypervisor big and prone to more erros though it will always be faster than the other two (theoretically).
XEN will not run Windows as fast as probably it will run on MS hypervisor. Running the primary domain as windows is more useful because all the hardware driver support and if windows run best on MS hypervisor then run MS hypervisor, run a minimal windows distribution which i believe will be 64-128 MB only with all the hardware drivers and service all other partition from this main partition.
What exactly do you think will make XEN succeed? Let aside the fact that management tools from MS and EMC will be superb and that is what enterprise need. Superb management tools which make managing all the virtual machines easy.
My feeling is that if MS pulls this one on time…it is going to take the glory…otherwise its EMC and XEN is probably the weakest link to bet on here…
>First of all i would like to say to all the OSS zealots that
>they if they feel that XEN will be the winner or its way
>cooler than other two technologies then they shouldn’t need
>to come and brag about it..correct?
No need for name calling, but I think you are refering to me. Still, I didn’t say Xen would win. I just said Xen is cool, free, doesn’t require hardware not yet available, and is available right now.
>Secondly, XEN the last time i checked was unstable as a 30
>year old car. No no don’t mess with it…it will crash..
Xen version 3 is out next month. It’s apparently used in enterprises. Next month I will probably use Xen to build a terminal server for the whole office.
What evidence do you have to say claim Xen is unstable?
>MS Hypervisor and XEN are almost seems like the same
>technology…run the hypervisor on bare hardware, run a
>domain or partition with device drivers for hardware and
>service other guest os from that domain.
In Xen, the domain 0 runs on the bare hardware and it includes drivers (such as disk and network). It seems like you think that’s two domains.
In Xen, I think it’s possible to expose additional hardware to specific domains. For example, you can grant video card access for GUI.
While there are few technical details available in this article, a clear difference is that Microsoft’s technology requires hardware not available until next year. Xen runs on current hardware, but Xen will also take advantage of the new hardware when it’s available.
>XEN will not run Windows as fast as probably it will run on
> MS hypervisor.
No one has enough information to claim that.
>Running the primary domain as windows is more useful
>because all the hardware driver support
What driver support? If you are running a headless server, all you need are network and disk drivers. Linux and NetBSD has great network and disk drivers.
>What exactly do you think will make XEN succeed?
Xen is already successful, it’s backed by academic and business resources, and it doesn’t need money to survive. (On the other hand, a lot of Microsoft products aren’t making money…)
Microsoft may make a good product, but it’s not here for a while. Again, my point is that Xen is cool, it’s free, and it’s already here.
Fox,
I noticed you are “passionate” about Microsoft’s unseen product. I also noticed your Washington state hostname: —-.hsd1.wa.comcast.net. What is your relationship with the Microsoft Corporation?
.hsd1.wa.comcast.net. What is your relationship with the Microsoft Corporation?
Why do you jump to conclusions that this guy is affiliated with Microsoft?. Real networks, Starbucks and many other companies also incorporated in .wa – there are a lot of people who live in Seattle that have absolutely no affiliations with MS.
Just because some one comes to the defense of MS, don’t call them a shill. Linux has its zealots and MS has its zealots and you just need to get along on this planet.
Why do you jump to conclusions that this guy is affiliated with Microsoft?. Real networks, Starbucks and many other companies also incorporated in .wa – there are a lot of people who live in Seattle that have absolutely no affiliations with MS.
It’s a question, not a conclusion. Maybe he will say he works for Comcast and I will look dumb.
Just because some one comes to the defense of MS, don’t call them a shill. Linux has its zealots and MS has its zealots and you just need to get along on this planet.
Fox did not “come ot the defense of MS.” No one is attacking Microsoft. The news.com.com.com article didn’t have a lot of information, so in the spirit of OSNews.com exploration and discussion, we were discussing. Then, Fox made some unjustified claims, which I hope we will discuss and sort out.
Thanks for the vocabulary word! For the record, Microsoft has been accused of (and I think caught) “shilling” before, but I doubt that today Microsoft would do it so obviously.
VMWare has the limitation that as new hardware comes, it will have to add new device drivers to its hypervisor, making hypervisor big and prone to more erros though it will always be faster than the other two (theoretically).
These benchmarks linked below shows Xen outperforming VMWare, and to their credibility, there are papers and independent verification.
http://www.cl.cam.ac.uk/Research/SRG/netos/xen/performance.html
I imagine that the VMWare developers are already working on taking advantage of the next generation hardware.
In terms of performance, it could be a close race between Xen, VMWare, and Microsoft. Notice in the benchmarks Xen is already running close to native speed. You can’t beat native speed.
First of all…yes i live in Washington. My relationship with Microsoft is hate x hate = love. Linux zealot hate Microsoft and i hate Linux zealots:) I was a Linux fan but not anymore. I lately started disliking Linux’s zealots attitude and their cursing Microsoft and any properietary stuff so now i like everyone except anything GPL…call me biased and yes i won’t say i am not.
Now lets come to some technical terms:
1. XEN doesn’t run unmodified OS so ofcourse you can get better performance.
2. I read XEN’s paper Safe hardware access with XEN and it talks about some DMA problems on why a device cannot be dedicated to a guest partition.
3. When i said XEN and MS hypervisor share same architecture…at least that what its look from the MS documents on the web. And I meant for VT based machines because i think the document says that microsoft hypervisor will only run on VT based machines. I know today XEN and domain0 runs in one partition and the reason is that there is no ring -1 on any x86 machine. When VT comes out i can’t think of a reason why XEN would run in domain0.
4. Virtualization is not only server business, client market is also hot. On clients driver support will matter a lot. Software development needs client virtualization.
5. Lastly i guess if microsoft makes windows and hypervisor it probably will run fast on their platform but i don’t know again its kind of a guess.
6. I talked about management software.you agree that is an essential requirement, right? who will make it for XEN?
Lastly being backed by academia or big corporations is not enough and you know that for Linux. Do you believe Linux is successful?
“Microsoft is no stranger to virtualization techniques. MS OS/2 1.x running on 286s used the virtualization techniques of the 286 chip, not not very efficiently – the 286 was not an efficient chip as regards virtualization. …”
This is a different class of virtualization.
With the proper hardware — and Intel/AMD are moving toward providing it — the virtualization allows strange events to occur; things that mainframes excel at such as dividing the processor(s) into chunks and each chunk is 100% of system speed and capacity. (Yes, this is simplified, though it’s not much off the mark.)
“Lastly being backed by academia or big corporations is not enough and you know that for Linux. Do you believe Linux is successful?”
I don’t think it’s possible to find many informed people who don’t think Linux is successful. Starting in 1991 with just a few guys hacking around and then getting to today with more marketshare than any other Unix/Unix-like OS in existence is pretty damned successful.
“First of all…yes i live in Washington. My relationship with Microsoft is hate x hate = love. Linux zealot hate Microsoft and i hate Linux zealots:) I was a Linux fan but not anymore. I lately started disliking Linux’s zealots attitude and their cursing Microsoft and any properietary stuff so now i like everyone except anything GPL…call me biased and yes i won’t say i am not.”
Er…so there are some @$$holes. They show up in every group. Here’s something to contemplate;
* What is the opposite of love?
* What is the opposite of hate?
The answer is the same for both questions.
Now, if you are engaged emotionally in either side of the debate, it’s because you are struggling to decide if you are indeed correct and have arrived at the conclusions you have reasonably. If you were not engaged emotionally — if you were apathetic — there would be no struggle. Correct or not, as far as you know there are no conflicts in what you say and what you truely believe.
As such, while you finger zealots for how they acted toward you and how irrationally they act in general, you are in turn close to the zealots you so dispise in your own actions.
If this is not true for you, so be it. I can only state what I observe.
Don’t assume that everyone is a zealot. Very few are; they are just vocal. Like yourself.
Um look what they did to VirtualPC. Enough said.
Let me lay out some concrete statements for arguments sake. Yes I work on Xen, so I have a vested interest. I’ll try to control my biases but feel free to call any of my points for more justification.
Xen:
Pros:
– Free (GPL). Not important to all but does at least mean you don’t have to pay cash to use it and it has the various other advantages (and disadvantages) of the GPL.
– Works now on vanilla x86
– Will eventually do full virtualisation and run Windows on VT / Pacifica hardware
– Ports of Linux 2.4, 2.6, FreeBSD 5.3, NetBSD 2.0, Plan 9 are available
– Market-leading performance (that’s not just hype – it really is very good)
– Supports just about any hardware Linux does
Cons:
– The current stable release can only run modified OSes
– Management tools are relatively immature
– Requires the use of Linux or NetBSD as a “host”
MS Hypervisor:
Pros:
– APIs for making guests VMM-aware => high performance possible
– Runs unmodified guests. If MS do it right, it’ll run Windows faster (at least initially) than the others
– Support from a big-name company
– Should have decent management tools
– Supports (just about) any hardware Windows does
Cons:
– Not available yet and will require future hardware (although it may be pretty common by then)
– Likely to be expensive
– requires the use of Longhorn as a “host”
VMWare ESX:
Pros:
– Runs on current hardware
– Rather high performance
– Mature product with good managemnent tools
– Run unmodified operating systems now
Cons:
– Very expensive
– Less hardware device support than the others
– Not quite as fast as running a VMM-aware OS
Some other points:
* We don’t know what the MS Hypervisor will look like by the time it’s released, so to a large extent discussion now can only be speculation
* Xen and VMWare have a big head start in features and quite conceivable will lead MS in features hen Longhorn Server is released
* Xen’s control tools have traditionally been their weak link. They’re getting there…
Not happening. The man said that the only thing OS X is going to be allowed to work on is an Apple computer.
Try getting support from Apple on your Xen box?
“Hi, my name is [username.XenMaster] and I have a problem with Mac OS X 10.5 [the Leopard, boys and girls]”.
“Sure, [usernamen.XenMaster] what seems to be the problem?”
“Well, I want to connect to the network, but it won’t let me”
“Ok, what kind of Mac do you have?”
“I don’t have a Mac, I’m running Leopard on a Xen box under simulation”
“I’m sorry sir, I’m afraid no Apple OS is licensed to run on any computer other than a Mac. We do not support any version of Mac OS X that does not run on a Mac. Thank you for calling.”
And that will be that. Afterwards there will be lots of bitching and moaning that Apple doesn’t support it’s users although it was perfectly clear from the outset that it was never going to be supported in any way, shape or form if the system doesn’t run on an Apple.
It’s going to run on an Intel anyway. Just buy a Mac. No more problems.
Yeah, I imagine running OS X on Xen on a non-Apple box will be in contravention of the license terms. There’ll probably also be technical measures in place to prevent this.
OTOH, running Xen *on* an Apple box might be possible and would allow you to run other OSes alongside MacOS X, with high performance.
That’d be a cool hack!
Virtualization is not only server business, client market is also hot. On clients driver support will matter a lot. Software development needs client virtualization.
Right now it seems VMWare is good at that, but what does it take? Easy setup or video support for guest operating systems?
I talked about management software.you agree that is an essential requirement, right? who will make it for XEN?
Essential for some, but probably not for many of the people using the currently supported operating systems. I plan to give Xen a try next month for our next server. In the documentation, the process seems easy.
The upcoming Xen 3.0 is supposed to offer:
Graphical control interface support. We need to develop a GUI for viewing and controlling resources across a cluster, either extending the existing web-based interface or developing a new GTK applicaiton.
and
Cluster management. We wish to add tool support to provide a better unified interface to managing a cluster of Xen machines. This would include features such as automatic use of VM live-migration to load-balance across a cluster, and the ability to ‘evacuate’ VMs from a node before shutting it down for maintenance etc.
Lastly being backed by academia or big corporations is not enough and you know that for Linux. Do you believe Linux is successful?
If you mean success is Linux taking over the world, no, but if you mean Linux fulfills a wide variety of needs in an efficient way, yes.
The Xen team are working on improving and robustifying the standard control tools. Simultaneously, a couple of groups are looking at a GUI node controller.
XenSource (the company started by Xen’s founders) are working on a cluster control tool for Xen. There are some really neat tricks a unified cluster controller, particularly with Live Migration.
And before I forget, OpenSolaris is being ported to Xen by Sun. I was under the impression they expected to have something working by now but I don’t know when the actual code release with Xen support will be.
Novell had apparently ported NetWare to Xen at some stage and Demoed it but I don’t know if they ever intend to actually sell that…
Now come on give us a break dude. Windows NT is a good operating system which is able to cater to needs of so many different variety of customers. If you really think that Windows is flawed then you probably should have given some concrete arguments. At the kernel level NT is more advanced than Linux with fully asynchrnous kernel, an awesome device model, full PNP, object ACL etc etc…Linux is just trying to come up there.
Will any of you (besides fox) comment on the accusations in this slashdot comment. I’m apt to believe any conspiracy theory about Microsoft lying, cheating, and stealing…
DRM (Score:5, Insightful)
by Frank T. Lofaro Jr. (142215) on Tuesday June 07, @05:12PM (#12751680)
(http://www.linux.com/)
They are doing this for DRM.
Their Hypervisor will enforce DRM, so even linux can’t override it.
They’ll make it so all device drivers must be signed to go into the Hypervisor which will be the only thing with any I/O privs that aren’t virtualized.
They’ll make it so new hardware has closed interfaces and can only be supported by a driver at the Hypervisor level.
Any drivers in any OS level won’t be able to circumvent the DRM, since they’ll just THINK they are talking to hardware, but will get virtual hardware instead – and the Hypervisor won’t let it read any protected content through the virtual I/O, it will blank it out (e.g. all zero bytes from the “soundcard”) or something similar.
The drivers designed for the Hypervisor won’t work in any higher level, since they’ll need to do a crypographic handshake with the hardware to verify it is “real” and the hardware will also monitor bus activity so it’ll know if any extraneous activity is occur (as it would if it was being virtualized).
Everything will have a standard interface to the O/S, so Linux will still run but be very limited and slowed down – since only Windows will be allowed “preferred” access to hardware, other O/S will be deliberately crippled.
They’ll say you can still run Linux.
Hardware manufacturers won’t release specs, they’ll say use the Hypervisor and you can still use Linux.
You’ll still need to buy Windows to use any hardware – Linux won’t even boot on the raw hardware.
MS doesn’t care if Linux isn’t killed – the above allows them lock in – no windows – your PC won’t boot – since nothing but the Hypervisor will know how to talk to the IDE card, etc.
What about manufacturers that want to support open interfaces, etc? Microsoft will deny them a key which they will need to talk to the Hypervisor – and the Hypervisor will refuse to talk to them.
Support anything other than solely the Hypervisor and you can’t use the Hypervisor. No Windows – lose too many sales.
And they can say other O/S’s are still allowed.
They’ll just not be able to give you freedom to use your hardware as you see fit (DRM, need to pay more to get software to unlock other features on your hardware), only Windows will run well, and you need a Windows license and Hypervisor for every PC or else it is unbootable.
This actually has me thinking about both Apple, MS and Sun:
Linux is the common enemy of Apple/MS/Sun so this is how they are going to cut off Linux’s air supply:
1) Apple will make sure that no machine besides Mac/Intel runs MacOSX. They’ll allow Windows to run on MacOSX but not Linux because Linux can be used to reverse engineer their bios/firmware/locks.
2) MS would like to have the rest of the x86 market all wrapped up in their DRM scheme and Intel and AMD are just too willing to see that happen.
3) Sun, through it’s partnership with MS and Apple will get the details. But Sun also has Sparc and their own AMD64 based machines so it may be safe.
4) IBM is on Sun’s, Apple’s and MS’s sh*t list and has exited the PC market and they can’t force intel/AMD to build DRM less machines. They are PPC shop now. They may have a license to build x86 clones but I don’t think they’ll go there. IBM has more fish to fry with XBox360/Sony/Nintendo
These events would force Linux/BSD/Hurd to the less popular and more expensive PowerPC, Sparc, MIPS, Itanium and legacy non-DRM’ed x86 platforms (incl. VIA/Crusoe).
Suppose that’s all true. Why would users walk into that? People will resist others restricting them. It doesn’t matter whether it’s running Linux or copying music.
Besides, I don’t see how this would work given economic interests (both vendors and consumers) and legal monopoly issues.
As a former BIOS tester, I have some insight you might want to hear;
“1) Apple will make sure that no machine besides Mac/Intel runs MacOSX. They’ll allow Windows to run on MacOSX but not Linux because Linux can be used to reverse engineer their bios/firmware/locks.”
Someone will find a way to run OSX on non-Apple hardware. They do it now on IBM Power systems, so not being able to do so on any other system seems unlikely.
Apple has stated they intend to make it possible to run either Windows or Linux on Apple hardware. They just won’t support it. How difficult it will be, though, is anyone’s guess. After all, you can run Linux on just about anything these days though it’s not necessarily trivial. (OSX is the software, btw. Apple Macintosh is the hardware.)
Additionally (from experience) you don’t need Linux to reverse engineer bios/firmware (while I don’t know what you mean by “locks”). In fact, it’s easier to use a stripped down system that ‘boots’ the BIOS and steps through that.
“2) MS would like to have the rest of the x86 market all wrapped up in their DRM scheme and Intel and AMD are just too willing to see that happen.”
This has no impact on Apple-brand x86 systems. Consider them otherwise alien.
As for DRM, it is destined to do two things;
1. Annoy customers.
2. Be defeated.
The reason is simple; A system not under physical control is not gauranteed to be secure. This usually applies for networked system security audits, though since DRM is a form of security it also applies to DRM.
“3) Sun, through it’s partnership with MS and Apple will get the details. But Sun also has Sparc and their own AMD64 based machines so it may be safe.”
Sun’s actions almost never make any sense. I’d like to shake some sense in them one day.
“4) IBM is on Sun’s, Apple’s and MS’s sh*t list and has exited the PC market and they can’t force intel/AMD to build DRM less machines. They are PPC shop now. They may have a license to build x86 clones but I don’t think they’ll go there. IBM has more fish to fry with XBox360/Sony/Nintendo”
Doesn’t matter. IBM, like Apple, gets it these days. IBM is interested in making money without killing off other companies (that could give them more money later). Apple’s move to x86 is pragmatic since all that matters is raw “horsepower”.
“These events would force Linux/BSD/Hurd to the less popular and more expensive PowerPC, Sparc, MIPS, Itanium and legacy non-DRM’ed x86 platforms (incl. VIA/Crusoe).”
Even without the comments I’ve made, I’m not sure what you’re talking about.