Xen High-Performance x86 Virtualization Released

Xen Team writes "The University of Cambridge Computer Laboratory Systems Research Group is pleased to announce the open source release of Xen, a virtual machine monitor for x86. Xen lets you run multiple operating system images at the same time on the same PC hardware, with unprecedented levels of performance and resource isolation. Even under the most demanding workloads the performance overhead is just a few percent --- considerably less than alternatives such as VMware Workstation and User Mode Linux. This makes Xen ideal for use in providing secure virtual hosting, or even just for running multiple OSes on a desktop machine."

The Xen team continues: "Xen requires guest operating systems to be ported to run over it. Crucially, only the kernel needs to be ported, and all user-level application binaries and libraries can run unmodified. We have a fully functional port of Linux 2.4.22 running over Xen, and regularly use it for running demanding applications like Apache, PostgreSQL and Mozilla. Any Linux distribution should run unmodified over the ported kernel. With assistance from Microsoft Research, we have a port of Windows XP to Xen nearly complete, and are planning a FreeBSD 4.8 port in the near future.

"Visit the project homepage to find out more, and download the project source code or the XenDemoCD, a bootable 'live iso' image that enables you to play with Xen/Linux 2.4 without needing to install it on your hard drive. The CD also contains full source code, build tools, and benchmarks. Our SOSP paper gives an overview of the design of Xen, and evaluates the performance against other virtualization techniques.

"Work on Xen is supported by UK EPSRC grant GR/S01894, Intel Research Cambridge, and Microsoft Research Cambridge via an Embedded XP IFP award."

Comments

Pfff

[AKAImBatman]

> Xen requires guest operating systems to be ported to run over it.

Get me all excited, then pull the rug out from under my why don't you? This is still pretty neat, but it's hardly a replacement for VMWare or Bochs.

Re:Pfff

[JohnFluxx]

I got excitted that windows would be ported to use it.. but then:

"The Windows XP port is nearly finished. It's running user space applications and is generally in pretty good shape thanks to some hard work by the team over the summer. Of course, there are issues with releasing this code to others. We should be able to release the source and binaries to anyone that has signed the Microsoft academic
source license, which these days has very reasonable terms. We are in discussions with Microsoft about the possibility of being able to make binary releases to a larger user community. Obviously, there are issues with product activation in this environment which need to be
thought through."

It would be a bitch if it was ported, worked perfectly, but then nobody was able to use it.

Re:Pfff

[rmohr02]

Only the kernel needs to be ported, and they have ported Linux 2.4.22 for you, and they're working on NetBSD, FreeBSD, and have almost finished Windows XP.

Re:Pfff

[phil reed]

VMWare still requires a port, doesn't it?. It's just that they provide a lot of them already done.

Re:Pfff

[Anonymous Coward]

No joke. If it were possible to run unmodified x86 OSes on this thing a million people would have just celebrated like the Cubs and the Red Sox were both in the series.

Free as in beer is still the reason I like open source.

Re:Pfff

[websensei]

read the article.

Xen only requires the OS *kernel* to be ported, not apps. For most users this is more than "good enough"... and makes it a viable replacement for vmware, especially given the reported (tiny) resource overhead.

Re:Pfff

[Spy Hunter]

Well, why not? It's definitely a replacement for user-mode linux, given the performance stats on their site, and it can replace VMWare for running Linux, if that's what you do with it. They say they have a port of Windows almost complete, but it'll be a cold day in hell before Microsoft lets them distribute it. So you can't run that last application that's tying you to Windows, or your games. But it can replace VMWare for quite a few other applications, or even play new roles that VMWare couldn't for performance reasons. (Remember that Z/OS virtualized Linux server thingy IBM was doing a while back? This could do the same thing only on Intel).

Re:Pfff

[jdh28]

No, VMware runs unmodified binaries.

What they do provide is ready made installations of various operating systems you can just install, although you have to pay for these (well, the MS ones at least).

Re:Pfff

[UberNerd]

No, you can run OS'es unmodified. But it provides *crappy* devices (an old 16 bit Soundblaster, AMD PCNet ethernet, an old, lesser-known SCSI interface).

Are you thinking of User Mode Linux? It requires the kernel to be patched, and UML provides a number of disk images (RedHat, Debian, etc) that will run under the patched kernel.

Re:Pfff

[Zocalo]

Not exactly. As far as I can tell from the quick scan of the PDFs and the statement that it just requires the kernel to be reconfigured, it's just a replacement hardware abstraction layer. I'm guessing that on NT/XP the only component that will need replacing is the HAL, and maybe some device drivers. If the concept takes off, then that's no different for a vendor than providing a different kernel/HAL tweaked for both single CPU and SMP systems, as many distros and Microsoft already do.

True, it's not a drop in like VMWare or Bochs, but this could be the killer app for people who run big VM based compilation servers like SourceForge. When you are talking iron of that kind of scale, a few percentage points saved on performance can easily equate to a significant amount of cash on the hardware budget.

Re:Pfff

[Anonymous Coward]

its good for open source OS's like Linux or *BSD, but won't be all that great for running Windows in a virtual machine unless Microsoft makes the kernel modifications available to everyone.

Considering that my main desire to use a VM is to run Linux, this is pretty nice and very useful.

Re:Pfff

[Anonymous Coward]

Except that beer generally isn't free.

I don't know who coined the ridiculous phrase "free as in beer" 'cause it just isn't true!

It is also a pretty stupid phrase as it involves trying to make yourself look cool by comparing things to beer, one of the fouler-tasting alcoholic beverages.

I suspect this usage derives from The Simpsons, where one of Homer's personality traits is an obsession with beer.

Of course, in The Simpsons this is carried off as quite funny, fitting in with the overall humour of the show. When the nerds parrot it, or adapt it to their own situations, the humour is quickly lost. Further indication of this may be found in the excessive quoting of other The Simpsons funnies on Slashdot.

In short, I don't like that phrase. At all.

Re:Pfff

[Deagol]

Yes, but VMWare still needs to make concessions for each supported OS. Yes, you can install and run Solaris/86, pre-v6 Netware versions, plan9, and BeOS under VMWare (with varying success), but they're not supported. They need to tweak VMWare for each OS. A true emulator, like bochs, doesn't need this tweaking.

I'm unsure why Xen needs guest OS mods, vs the way VMWare and Plex86 do things. I only skimmed the paper, so I might have missed something.

I, too, was excited until I read the "porting" part. I guess free Linux and *BSD virtualization is better than nothing. But it'll be nowhere near as comprehensive as VMWare (too bad -- I want to ditch the over-priced beast) and plex86. However, kudos to the Xen team, as it really is a cool project.

Isn't IBM porting their hardware partitioning stuff to linux, too?

Re:Pfff

[spektr]

The Windows XP port is nearly finished. It's running user space applications and is generally in pretty good shape thanks to some hard work by the team over the summer.

I'm surprised that this was possible at all with Windows XP. Microsoft's Shared Source program doesn't seem to be as useless as I thought.

Of course, there are issues with releasing this code to others. We should be able to release the source and binaries to anyone that has signed the Microsoft academic source license, which these days has very reasonable terms. We are in
discussions with Microsoft about the possibility of being able to make
binary releases to a larger user community.

I think there are two possibilities to do this. First, Microsoft incorporates the changes into their main trunk or releases patches for it. Second, Microsoft allowes a group of hackers to distribute modified Windows binaries. Both alternatives don't seem very probable.

Re:Pfff

[kalidasa]

MS is one of their sponsors. How much you want to bet that MS licenses the technology and distributes it under their own logo? We already know from the VirtualPC purchase that this is a future direction for them.

Re:Pfff

[AKAImBatman]

Not quite true. Say I want to grab the latest Knoppix or Red Hat and do some testing of my application. I'd have to build my own modified distro of these in order to run them. VMWare and Bochs allow me to simply create a virtual hard disk, point to the ISO, and install. And trust me, all distros are NOT created equal.

Re:Pfff

[kzinti]

In being so quick to chastise the guy for not reading the article, you missed the point. If anything at all needs to be ported, it means that an OS like XP can't be installed on Xen out of the box. It means that a port of the XP kernel has to be completed, and that MS has to authorize the ported kernel to be distributed. For lots of folks, this is not good enough... even assuming that MS ever would allow the ported kernel to be distributed (and then we have to worry about activation issues too). So, while Xen is great for allowing linux users to run multiple distros at max performance at the same time, it doesn't much help someone who wants to run XP, and might never help them. Those folks still have to resort to VMWare or Bochs.

Pfff...

Re:Pfff

[Anonymous Coward]

Activation isn't handled by the kernel -- it's part of the Win32 subsystem and is entirely user-mode.

Re:Pfff

[quigonn]

A true emulator, like bochs, doesn't need this tweaking.

Then, tell me, why I can't get Coherent 4.2.10 to run with Bochs, because Coherent thinks the floppy drive is open?

Re:Pfff

[Anonymous Coward]

Still means I can't use my existing copies of Windows 98-XP. This means I'll have to spend much more money to do the same things in a VM which is free. Unless the ported kernels still run on a machine without Xen, in which case it could be useful.

I want full DX9-ability!!!!

Re:Pfff

[larry bagina]

How about "free as in herpes". It fits GPL to a T - it's viral, you can share it, but you can't get rid of it once you're infected.

Re:Pfff

[kzinti]

Activation isn't handled by the kernel -- it's part of the Win32 subsystem and is entirely user-mode.

So what? The problem with activation is how it identifies the machine. I install XP on partition A. I install linux on partition B. Then on partition B, I use Xen to make a second installation of XP, or I try to run the version install on partition A. What matters to me as a user is whether the copy of XP running under Xen needs a second license. If the activation algorithm thinks that both copies of XP are running on the same machine (which they are), then I can activate both copies of XP under the same license. If not, I have to buy a second copy - even though I'm running it on the same machine.

Re:Pfff

[Anonymous Coward]

I really don't get people's obsession with beer either. It tastes awful -- when I hear someone say one beer tastes better than another, it reminds me of our political system where you get to choose the less evil of the two scum sucking bureaucrats. All beer tastes like piss, just some are slightly more tolerable than others (and yet this topic sparks endless, pointless debates).

I suspect that people are rallying their approval of the effects of the alcohol rather than the beer itself. It just so happens that beer is relativiely inexpensive and readily available.

Reading between the lines, I suspect most of beer's advocates are too self-conscious to be able to interact with unfamiliar people who they find sexually attractive without being inebriated. So basically, in the end, people probably see beer as a cheap lottery ticket to the opportunity to engage in sex.

Sex is the only reason I can see why people would want to consume and maintain the popularity of such an awful-tasting beverage.

Re:Pfff

[homerules]

My favorite beer is coldandfree

Re:Pfff

[Deagol]

Let me re-phrase: A true emulator, like bochs, shouldn't need this tweaking.

When I last read up on bochs, he stated he was shooting for 100% IBM-PC emulation. If acheived, my original would be true.

Re:Pfff

[frp001]

The expression "free beer" comes from the free software definition. And if my memory is correct, it also outdates the Simpsons.

Re:Pfff

[homerules]

It's clear to me that people who are around you need to drink beer in order to tolerate your whining. Most of beers advocates can control their drinking and consume for pleasure. maybe if you did some hard work in your life you would understand the pleasure of a cold beer after a hard days work.

Re:Pfff

[vrt3]

Yes, but VMWare still needs to make concessions for each supported OS. Yes, you can install and run Solaris/86, pre-v6 Netware versions, plan9, and BeOS under VMWare (with varying success), but they're not supported. They need to tweak VMWare for each OS.

I could be wrong, but I didn't think so. Not supported means they don't support it, not that it doesn't work. The tweaking that is done when you select a specific OS only concerns the configuration, such as amount of memory and stuff. Things you can easily adjust yourself.

I'm unsure why Xen needs guest OS mods, vs the way VMWare and Plex86 do things. I only skimmed the paper, so I might have missed something.

VMWare and Plex86 need to trap priviledged instructions in running code and replace them with their own. I believe they do this by literally scanning code-to-be-executed, which slows down the virtual machine of course.
Xen solves this by replacing all such instructions with Xen system calls; somewhat the same as User Mode Linux, where low level instructions are replaced with system calls to the host kernel.

Re:Pfff

[Spy Hunter]

I don't see what this technology can do for Microsoft. They don't want you to be able to run Windows on Linux, because that would make it easier to switch to Linux. They don't want you running Linux on Windows because they'd rather you just ran Windows. Xen can't run Windows on Mac OS, or vice versa. The only way I could see Microsoft using this technology is as a way to run multiple copies of Windows at the same time. That might be useful for developers or people who want sandbox systems for security reasons, but it seems to me like these uses are much less compelling. Xen just makes more sense as a way to run multiple different operating systems at the same time, which is exactly what Microsoft doesn't want.

Re:Pfff

[beowulf_26]

Blashphemy!!!! Beer is the nectar of the Gods!!

Re:Pfff

[coopaq]

Xen requires guest operating systems to be ported to run over it. This article just became completely useless to 99% of readers. What everyone wants is a free x86 virtual machine running on Linux, Windows or FreeBSD. Good idea, but no thank you. Next.

Re:Pfff

[Anonymous Coward]

Actually that phrase is only part of the equation and I assure you that you like that part much better than the alternative.

I don't know the origin, but it goes something like this :

(Couple of guys discussing something new) ... middle of the discussion:

Guy1 : ... and it is FREE!
Guy2 : Free as in beer, or free as in herpes?

See free beer is something you would like, but ...

Re:Pfff

[beezly]

The license used to enable them to do this modification isn't the Shared Source License, it's the Microsoft Academic Source License. It's completely different.

Re:Pfff

[Dashing Leech]

You seem to have missed the fact that different people have different tastes. I happen to love the taste of (some) beer. But I can't stand the taste of coffee, and don't like most wines.

The difference is that I recognize my tastes don't apply to everyone, and I certainly don't try to interpret common usage of things dislike in terms of my own tastes, like you did when you said "I suspect that people are rallying their approval of the effects of the alcohol rather than the beer itself." Did it occur to you that other people might just like the taste of beer, even though you don't.

All you coffee drinkers are alike, you generalize about everything. (=

Re:Pfff

[alexdewaal]

/One of the fouler-tasting alcoholic beverages/ It all depends on the kind you refer to. If you are referring to the best selling US variants, I'm with you. However, there are numerous variants that are quite good and a smaller amount simply delicious. Get Informed! http://www.beerhunter.com/documents/19133-001726.h tml Is that alcohol-free? No, but it's free alcohol!

Re:Pfff

[Fluffy the Cat]

I'm surprised that this was possible at all with Windows XP. Microsoft's Shared Source program doesn't seem to be as useless as I thought.

Microsoft Research UK is literally next door to the University of Cambridge Computer Laboratory. Microsoft Research UK desperately wants the university to like it in order to gain more academic credability.

Re:Pfff

[Rick the Red]

Bochs isn't finished. Give the guy a break. Either help him finish it or don't bitch because it won't run your pet obscure OS. Would you rather open source projects wait until 1.0 to release their code?

Re:Pfff

[KrispyKringle]

They're not working on FreeBSD.

Unfortunately, the NetBSD port has stalled due to lack of man power. We believe most of the hard stuff has already been done, and are hoping to get the ball rolling again soon. In hindsight, a FreeBSD 4 port might have been more useful to the community. Any volunteers? :-)

Re:Pfff

[Martin Blank]

I could be wrong, but I didn't think so. Not supported means they don't support it, not that it doesn't work. The tweaking that is done when you select a specific OS only concerns the configuration, such as amount of memory and stuff. Things you can easily adjust yourself.

Correct. Debian's not supported, but it works. OpenBSD isn't supported, but it (sort of) works. Basically, you have to be prepared to deal with some frustration in things being slow or sometimes unstable, but this does not have to do so much with the emulation as it does with the tools. VMWare has a package that is installed on the virtual machines that contains drivers for the hardware that does need to be virtualized, and a couple of utilities allowing access to a "shared drive" to allow easier movement of files between main system and VM. The Linux tools, BTW, can be compiled individually for each OS. Once I finally got Debian installed, this made things much smoother.

Re:Pfff

[kma]

They need to tweak VMWare for each OS. A true emulator, like bochs, doesn't need this tweaking.

How high are you? Do you really think VMware, Inc. pays me to "tweak" our monitor so ~12 slashdot-reading customers can try running plan9 in a VM? Plan9, OpenStep, AtheOS and the rest of the hobbyist OS'es just work, because VMware contains a basically correct x86 virtual machine monitor. Much like they just work on bochs, a basically correct x86 emulator. The VMM and the emulator accomplish the same computational task; they simply use different techniques to accomplish it.

Re:Pfff

[Haeleth]

Given that Bochs is already past 2.0, merely waiting till 1.0 for a release probably wouldn't solve much.

Re:Pfff

[kscguru]

I'm unsure why Xen needs guest OS mods, vs the way VMWare and Plex86 do things. I only skimmed the paper, so I might have missed something.

VMWare and Plex86 need to trap priviledged instructions in running code and replace them with their own. I believe they do this by literally scanning code-to-be-executed, which slows down the virtual machine of course. Xen solves this by replacing all such instructions with Xen system calls; somewhat the same as User Mode Linux, where low level instructions are replaced with system calls to the host kernel.

Not quite ... they just run the virtualized kernel in plain user-mode, then when a privilaged instruction comes up it trips a processor exception. The VMM (VMware or Plex86) then steps in, emulates the privilaged instruction, then gets out of the way. Privilaged instructions, for the non OS-enthusiasts, are usually hardware-related things like INP and OUTP, instructions to change the page tables, switch between 8086 and protected mode, etc. Since these instructions are so rare, VMWare is only ~10% slower than normal - and runs full speed on a large set of workloads.

Not having read the Xen site yet, what they _probably_ do is replace all such instructions in the source code with their own function calls. VMware did something similar at one point, but didn't turn it into a product because they felt it didn't have commercial potential.

Re:Pfff

[TheLink]

Actually the O/S selection stuff seems to be more for the VMWare tools.

If you select windows as the OS when you choose the option to install Vmware tools the "CD" for Vmware tools for windows is mounted. Whereas if you selected Linux, a different "CD" would be mounted.

The tools are optional. You don't need them to run the O/S. They do stuff like allow copy and paste between host and guest O/S, host O/S folders to be mounted as drives, mouse pointer seamlessly moving between host and guest, better optimized video drivers. That sort of thing.

Performance seems to be fine as long as you have enough RAM.

Re:Pfff

[Deagol]

A statement I read on on of the VMWare newsgroups led me to the conclusion of mine you quote above. Someone was bitching "why can't XYZ OS run under VMWare if it runs perfectly well under generic PC hardware". Someone responded to the effect that VMWare was tuned to make certain concessions for the OSs they supported. My foggy memory tells me it was an actual VMWare employeer.

Though many unsupported OSs do in fact work (I have all 3 of the free BSDs running at once in VMWare, though only FreeBSD is officially supported), some just don't work. Back in VMWare 2.x days, I could get one version of Solaris/86 to work, but not another.

Maybe it is the guest OS's fault for not handing the virutal hardware correctly, I don't know for sure. But it's by no means a guarantee that all PC-based OS's will "just work" under VMWare.

Re:Pfff

[cookd]

1. Drivers. 'Nuff sed.
2. Performance. If they made the emulation perfect, everything would run perfectly, but there would be an unacceptably big hit on perf. Instead, they can adapt their emulation according to what the OS can tolerate reliably.

Re:Pfff

[kalidasa]

Work on Xen is supported by UK EPSRC grant GR/S01894, Intel Research, and Microsoft Research.

Why do you think MS would be supporting this work financially if the only conceivable uses for it are one that are "exactly what Microsoft doesn't want?"

Re:Pfff

[christophersaul]

Everyone has different tastbuds, but I think beer's lovely. Men tend to be bigger fans than women... No other alcoholic drink provides the mix of a feeling of wellbeing, happiness and general downright comfort than beer does!

Re:Pfff

[kma]

Someone was bitching "why can't XYZ OS run under VMWare if it runs perfectly well under generic PC hardware". Someone responded to the effect that VMWare was tuned to make certain concessions for the OSs they supported.

Right. The OS'es we "support." Not the OS'es "we're able to run." And those "concessions" are mostly performance trade-offs, not correctness.

It's still true that some OS'es don't run; but that's because our software has bugs. OS'es sometimes have bugs, too, though if the OS is important enough, we'll work around it. The bugs that prevent you from running the OS that some drunken Swede cooked up for course credit are admittedly less important to fix than the bugs that, say, prevent you from running Linux. However, in the long run, we try to fix even the bugs exposed only by drunken Swedes.

That's why AtheOS, OpenStep, BeOS, NetBSD, FreeDOS, B-Right, Plan9, QNX, and myriad other commercially unimportant OS'es run ok. Not because they're important to our customers (man, oh man, they aren't), but because they enable us to be sure our x86 virtualization layer is reasonably correct. That way, when Ingo Molnar decides to start using 80286-style 16-bit tasks with lots of grow-down and conforming code segments to do system calls in Linux 2.6, we won't get caught out too badly.

Re:Pfff

[Torne]

You need a lot more than a replacement HAL unfortunately. Under Linux the changes are relatively minor as the x86-specific stuff is confined to the x86 arch tree, and very little of the non-arch code needs modification to run under Xen. Windows, however, assumes x86 all over the place, and we've had to modify the HAL, kernel, and a few system libraries so far.

Torne
(XenoXP porter)

Re:Pfff

[rifter]

I really don't get people's obsession with beer either. It tastes awful -- when I hear someone say one beer tastes better than another, it reminds me of our political system where you get to choose the less evil of the two scum sucking bureaucrats. All beer tastes like piss, just some are slightly more tolerable than others (and yet this topic sparks endless, pointless debates).

I used to think this, but then I tried non-American beers. For some reason US breweries seem to think that beer is supposed to taste the same before and after it is run through the body. Wih the rise of microbreweries, this has been changing. But the big breweries still make piss-beer.

Then again, probably much like piss, you can accquire a taste even for the bad beer if you really, really try. Free beer is often the piss tasting variety because people generally do not give away much of the expensive kind.

Re:Pfff

[rifter]

Except that beer generally isn't free.

I don't know who coined the ridiculous phrase "free as in beer" 'cause it just isn't true!

Maybe for you because no one likes you. Seriously free beer is pretty damn common, especially at parties. It is not free for the host, this is true, but it is free (gratis).

There are also all kinds of events where you can get free beer because free beer is a good way to attract people. Aso there are tours of breweries which are often free and come with many "samples."

Re:Pfff

[vrt3]

A few years ago, I loosely followed Plex86 development. One of the things I saw mentioned on the mailing list was something about scanning; Google turned up with http://systems.cs.colorado.edu/ComputerArchitectur e/Bochs/plex86/PIG/c10.html. That's were I got the idea about the code scanning. I didn't study any of this in any detail, so you might very well be right.

Re:Pfff

[Anonymous Coward]

No, you're wrong. Not all instructions that should be privileged are on the x86. There are several instructions which allow a user-mode program to detect that it is indeed a user mode program, and not supervisor mode, without signalling any kind of trap or exception. For example, you can look at some of the upper bits of the EFLAGS register or do an RMSW to access the low 16 bits of CR0. So you do have to scan for such "bad" instructions, and interrupt the code whenever it's about to get past the end of your scan, to continue the scan. One way to do this is to set breakpoints using the debug registers (DR0-3, and maybe 4-7?), another is to replace instructions (but you have to make the code page unreadable then, and I believe all x86 executable code pages are also readable); you can do this modification either in-line, or in a separate trace cache, and use the trace cache to speed up code, hopefully compensating for loss due to virtualization (much like the Crusoe and Pentium 4). So there.

Re:Pfff

[Anonymous Coward]

So there.

this was the best part!

Re:Pfff

[Spy Hunter]

Because Microsoft is a humongous company and the Research branch (which is the part of Microsoft doing the sponsoring) is a completely separate entity from the Windows branch. When I say this is "exactly what Microsoft doesn't want", I am talking specifically about the Windows and Office parts of Microsoft, and not the Research, MSN, XBox, hardware, or Expedia branches, or any other branch. These branches are all run by different people who may not always be working in perfect harmony. It's kinda like the fallacy of the single Slashdot poster: People often assume that everyone on Slashdot shares the same opinion about everything, which is not true. Likewise, people often assume that companies are single entities that always act in a coordinated, logical manner. Instead companies (especially large ones) are a mishmash of smaller departments, each governed by a different set of people, who may have different opinions and goals which may conflict.

Re:Pfff

[kscguru]

Not wrong, just glossing over the minor details :-).

The VMware guys did complain about the two or three (he was deliberately vague) instructions that behave wrong for virtualization - the x86 design is, in fact, buggy enough to make virtualization difficult.

This only matters when the OS uses, and cares about, the particular instructions. I think earlier versions of VMware ran slightly modified kernels with these instructions directly changed, until they got patches into the respective entities (e.g. Linus) to eliminate the rare instruction that misbehaves - the replacement is only a handful of instructions anyway, and only in hand-coded assembly because no compiler uses such instructions, so it makes no performance difference.

Not 100% pure virtualization, where the virutal OS can't detect it's being emulated, but there are so many other ways to check ... fire off a bunch of TLB instructions, for example, and time it against the RTC, a VMM will take much longer due to emulation. But it would be stupid to try to fake out an OS anyway since it's so easy to detect!

Hmm... now that I think about it, the method I've described was mentioned specifically for VMware's attempts to run VMware inside of VMware - I'm not sure that they don't do code rewriting in normal use, but they CAN do complete virtualization with practically no changes to the underlying OS - including no dynamic code rewriting!

Re:Pfff

[rmohr02]

Yea. I re-read it and noticed that. But they are trying to get people who can work on it.

Re:Pfff

[weileong]

Actually, I've found that VMware doesn't work with quite a lot of the x86 OS-es I've tried (and on their website there's something to that effect).

BeOS doesn't work, neither does QNX or OS/2. I'm having a heck of a time getting Solaris x86 to work as well (doesn't seem to like the HDD being presented as a SCSI device, must be IDE, although I've installed actual Solaris x86 machines with SCSI HDDs before without problems).

Re:Pfff

[Anonymous Coward]

VMware tweaks performance for each guest OS, but the overall system is almost entirely general.

I tell VMware I'm running Linux when I use it to test/debug my homebrew kernel.

Of course, it helps that I'm employed by VMware and can modify their monitor to help debug my kernel.

Not speaking for VMware, Inc, of course.

Re:Pfff

[Anonymous Coward]

Have you ever excuted an SIDT instruction in a VMware virtual machine (presumably plex86 and Microsoft Virutal* are the same).

You get their virtual machine monitor's IDT, not the one you expect.

Apparently (given that no-one bothers to virtual this instruction) it doesn't affect most mainstream guest OSes.

Re:Pfff

[Anonymous Coward]

VMware successfully virtualizes all instructions. You can run unmodified linux 0.0x if you like.

Recent versions of plex86 rely of modifying the guest OS to not use dangerous operations. VMware goes to enormous lengths to virtualize these 'difficult' operations.

Re:Pfff

[Anonymous Coward]

BeOS mostly works for me. I've only tried it in OBJ (read: devel) builds of VMware, but I have no reason to suspect it won't work in Workstation 4.0.

The clock moves at 1minute/second or so, but one of my coworkers (who was on the kernel team at Be, Inc) assures me that that is (mostly) a BeOS bug.

Re:Pfff

[Anonymous Coward]

more patchset, less arrogant prancing on useles web-boards.

come on now.

Re:Pfff

[Anonymous Coward]

Microsoft Research UK desperately wants [Cambridge University] to like it in order to gain more academic credability.

As someone with connections to the Computer Laboratory, I can promise anyone wondering that the above will happen when Microsoft demonstrate R&D worthy of academic credibility, and not because they built a big building or donated significant amounts of kit. Donations are gratefully received, of course, but they won't make anyone in the lab think that Windows XP is a secure operating system or Visual Basic is the programming language of the future.

Re:Pfff

[weileong]

Hrm, the installation just bombs on me - VMWare pops up an error dialog and I can't get any further.

So, you're a secret internal VMWare guy eh... how's working for the company like? You guys worried about MS-and-Connectix at all?

Re:Pfff

[Anonymous Coward]

First, Microsoft incorporates the changes into their main trunk or releases patches for it.

Well, this isn't as far-fetched as you might suppose. Afterall, MS is currently building a VM for server consolidation -- why wouldn't they want to make it 4 times faster under server workloads?

Funny Guys

[tomblackwell]

... to publish a paper called "Xen And The Art of Virtualization".

Re:Funny Guys

[kzinti]

...and that paper title tells us how to pronounce Xen, at least if you're familiar with the Robert Pirsig title.

How is it ideal for secure virtual hosting?

[Anonymous Coward]

What qualities of a VM solution necessarily translate into security?

Re:How is it ideal for secure virtual hosting?

[Zigg]

Give each customer their own machine that they can do pretty much what they please with, root rights and all. No customer gets rights over the other customers' VMs, though.

Re:How is it ideal for secure virtual hosting?

[Anonymous Coward]

It's all about (the illusion of) control ..

Re:How is it ideal for secure virtual hosting?

[__past__]

Virtualization rocks for that kind of things. I happily use FreeBSD jails for that. They work on a different level and have different goals, but giving each (group of) user(s) their own sandbox to play with is definitly a really cool thing, both for admins and for the users themselves.

Re:How is it ideal for secure virtual hosting?

[Zigg]

chroot is not a complete solution. You may be able to restrict people from getting at files above their chroot level, but there a host of other things a root user can do to the system that you don't want your customers to be able to do if they can affect other customers.

Re:How is it ideal for secure virtual hosting?

[Tuqui]

root matrix?

Re:How is it ideal for secure virtual hosting?

[bill_mcgonigle]

Hopefully, this will help drive down the cost of hosting. I need my own mod_perl, so I need my own apache, so I need my own linux, so I need my own hardware.

Now, we can drop that last clause and 10 users can share a box, so maybe I can get my own host for $20 a month instead of $100.

Re:I bet it's not Open Source...

[Zigg]

RTFA. It's GPL.

(waiting for the 20 second delay to expire... @#%$! Slashdot)

oh yes yes yes yes yes

[moodz]

I'll have what they're having !! ... well at least as soon as I can get hold of it.

Re:I bet it's not Open Source...

[Trigun]

With assistance from Microsoft Research, we have a port of Windows XP to Xen nearly complete

They won't release the source for XP, but you can probably get a compiled binary. I just wonder if you'll have to re-register every time you change your virtual hardware.

MOL for x86?

[Realistic_Dragon]

In the same way that Mac On Linux makes moving people to Linux on a Mac pretty painless (just give them an icon for OSX in a window) this might do the same for migrating Windows users.

The biggest problem with emulators/virtualisation has always been speed. If a system can be set up that runs Linux but can boot XP easily and run fast, that will be a big improvement.

Of course it's not going to be much good for gamers (doesn't look like it can use hardware accelaration) but it's still pretty promising.

Re:MOL for x86?

[Anonymous Coward]

More like it allows people to run the two side by side and realise that on the Mac OS X is far superior to Linux as you can do so much more with it ..

Re:MOL for x86?

[tgd]

This would not allow you to run Windows under Linux... it would allow you to run Windows and Linux under Xen, which is not nearly the same thing.

Contrary to the submitter's comments, this product is nothing like VMWare.

Re:MOL for x86?

[BigJimSlade]

it would allow you to run Windows and Linux under Xen, which is not nearly the same thing.

If it's true that Xen is the base and doesn't sit on a host OS (I can't RTFA, because it appears to be slashdotted at the moment,) it would be very much like VMWare's ESX Server. This is different from their GSX Server in that it is practically its own OS, and is installed on the bare metal of the box.

Re:MOL for x86?

[Glonoinha]

Actually based on my skimming the article it merely removes the requirement for the host OS in VMware - Xen sort of creates a (Null) host OS from which virtual machines are started.

In that respect it is pretty much like a version of VMware that doesn't require a host OS. Oh yea - and it doesn't actually run any commercially available OS without modifying and recompiling.

It looks cool, no doubt, and a LOT faster in certain aspects where multiple layers of abstraction are slowing down VMware (according to the nifty graphs, I haven't actually tried it) out on the fringes of performance.

That said, I really like VMware. Like the second coming of the silicon jesus - it is good. Wicked fast compared to some other VM implementations I have run (does Desqview on a 386 count?) Anybody that is reading this and hasn't played with it needs to get the 30 day trial from www.vmware.com

Imagine all the stuff described in this post, but it works with the OSs you already have, requires a host OS (needs to be run under Linux or Windows) and cost $300.

Re:MOL for x86?

[Anonymous Coward]

Mac On Linux makes moving people to Linux on a Mac pretty painless

why move when you can have both? and at the same time.

also, if we consider OS X (and not OS 9) we are talking about moving from a Unix system to a Unix system. So it is more a matter of personal preference. There is not much gain for a non-developer except for speed. But if KDE or other fancy WM is used, then the speed advantage might go away.

I'm using GNU/Linux most of the time on my iBook because:
1) it's faster
2) I am more familiar with that development platform as I was working on Debian GNU/Linux/x86 before. I just switch hardware...

But for a "real" Mac user, except for 1) there is no real advantage.

Re:MOL for x86?

[Virtex]

Ok, so this is a nanokernel. Something similar to Adeos.

Re:MOL for x86?

[Anonymous Coward]

You may want to look at the VMWare ESX server product. Does this, works, is being heavily supported by the tier one players, doesn't require munging up operating systems.

Re:MOL for x86?

[Anonymous Coward]

"I'm using GNU/Linux most of the time ..."

It's just Linux.

Unless you run perl/python/php/ruby/apache/mysql/postgres/sendmai l/exim/postfix/qmail/bind/djbdns/BSD/GNU/Linux.

Yet another reason to run FreeBSD ... :>

Re:MOL for x86?

[Glonoinha]

Oh man in today's economy I had a hard time coming up with $300 for regular ol' VMware. I think ESX is listed on their price sheet as 'If you have to ask, you can't afford it.'

Which today is pretty accurate, in my case.

Re:MOL for x86?

[Torne]

Yes, this is exactly the case, except ESX emulates a full x86, whereas Xen exposes our ring-1 architecture, xeno-x86, and thus guest OSes have to be ported. Xen is much faster than ESX as there is no need to emulate privileged accesses.

Torne
(XenoXP porter)

Re:MOL for x86?

[Kris_J]

I want to run Linux under Windows, is there a product to do this? A single file on my Windows PC with the entire Linux OS in it, ssh access as standard with the ability to install VNC for X would be ideal.

Re:MOL for x86?

[Anonymous Coward]

Wrong.

Re:MOL for x86?

[Anonymous Coward]

uhh, ESX doesn't emulate privleged instructions, it BTs them.

Re:MOL for x86?

[Anonymous Coward]

Vmware.

Interesting ..

[Anonymous Coward]

I'm really surprised they got assistance for windows XP. You'd think that Microsoft would want to stay well clear of anything linked to the GPL.

I guess such assumptions are incorrect, and quite probably a result of reading Slashdot too much!

Re:Interesting ..

[sjwt]

I dont know about it being becase you read too much /. , but i think the reson they would be willing to help is that this would look realy good in any future AnitTrust case..

"your honnor, we have helped the open source comunity devolpe a program that meens anyone can run any OS along side MS with out any problems, we dont see how you can accuse us of being Anti competive when we allow others to compeat on the same machene"

could you even imagen a futer MS install CD where it automaticly installs a VM *inx?

Re:Interesting ..

[Gingko]

This is Cambridge computer lab we're talking about - where Microsoft research have the very next building over.

I live with one of the guys in the systems group and hope to go back there to do my PhD soon, and they do do very cool things there. Microsoft give them all the help they need, because academia is an excellent ideas feeder for the real world. Cambridge, being in a position of power with its serious reputation and fantastic set of minds, gets the benefit of the Microsoft help without any of the assumed costs.

Henry

Re:Interesting ..

[AArmadillo]

Its fairly wierd actually. Microsoft also has a program where if a group of college students want to work on some computer science/engineering related project but do not have the necessary resources, they will help provide funding for it -- even if it is an open source or GPL'ed project. I suppose Microsoft's help with Xen is an extension of this program. I don't think their interests are totally philanthropic either -- if Xen happens to hit it off and become widespread, they're definately going to want it to run Windows.

Re:Interesting ..

[Anonymous Coward]

I highly doubt that that is the case. It seems to me their motivation would be more technological than legal. Sounds like you've been reading too much Slashdot too!

Re:Interesting ..

[urdak]

Well, let's see: apparently Xen lets you run hyped Linux applications (such as Apache) on your Windows machine, letting you stay with the operating system you're addicted to.

At the same time, if you're a Linux user and want to run Windows applications, you're screwed because you'll have to wait for Microsoft to release the Xen-patched version of Windows XP. And guess what pigs will do when that happens?

So, tell me again why this Xen doesn't *exectly* fit Microsoft's usual M.O. ?

Re:Interesting ..

[Anonymous Coward]

It's slightly less weird than you may think. Microsoft does a lot of research in house (they're kind of like what bell labs used to be in the good old days) and provides grants for student research. I was involved in a group that had a research grant. It's a GREAT way for microsoft to find and recruit future talent.

Re:Interesting ..

[neil_rickards]

I already run Apache on my Win2k box and it works like a beut!

Re:Interesting ..

[Anonymous Coward]

well, it's good that you admire us, but, really, you are a fawner and nobody likes you here. If we took people like you, we wouldn't be the "fantastic set of minds". And internally, we run linux. We just work on XP part-time for the funds.

Paul

Re:Interesting ..

[Anonymous Coward]

how very charming. Always interesting to see who takes the bait ;)

Quite why you take such offence at an objective observation that people at the computer lab are generally pretty bright is beyond me, but as you so delicately pointed out, it's probably just a function of my inferior intellect. Yes, other people are also clever, well done. As your post demonstrates, the signal:noise ratio is a little lower than I had hoped. Either that or you were laughed at so much as a child that your ability to take a compliment is heavily impaired.

The linux / windows debate got very old very quickly. I certainly don't give a flying fuck what you use, but it's fair to point out the close connection between MS and the Lab (Roger Needham, for example?). But, y'know, well done for making yourself sound like the worst kind of sell-out.

Can't help but smiling at the "if we took people like you" sentence. Firstly, identifying an arsehole like you with the entire admissions procedure is a bit of a stretch. Secondly, uh, 'you' already did. But thanks for playing!

jeez.

Re:Interesting ..

[Curt Cox]

Cambridge, being in a position of power with its serious reputation and fantastic set of minds, gets the benefit of the Microsoft help without any of the assumed costs

Some consider the loss of one's soul to Satan a cost.

Re:Interesting ..

[Anonymous Coward]

Yeah, because Haskell and F# are taking the world by storm. Exactly what other ideas has Microsoft Research Cambridge fed to "the real world"?

Re:Interesting ..

[Anonymous Coward]

Ok. I apologize.

I thought you was a marketing droid.
I'm sorry.

Not really like VMWare

[javatips]

We have a fully functional port of Linux 2.4 running over Xen

With assistance from Microsoft Research, we have a port of Windows XP to Xen nearly complete, and are planning a FreeBSD 4.8 port in the near future (volunteers welcome!).

If one need to port an OS to make it work within Xen, then I will NOT compare it to VMWare. VMare can run your stock OS on a VM whithout the need to tweak it.

The performance advantage it has over VMWare is probably related to that. By having a few restriction on the OS, they can probably offer better performances.

Re:Not really like VMWare

[rmohr02]

They've ported Linux 2.4.22 and are nearly done with a port of Windows XP. If it's relatively easy to change these, I will definitely try it.

Re:Not really like VMWare

[Anonymous Coward]

No, it is like VMWare in that several different virtual operating systems can be run on the same machine concurrently. They differ in method, but have essentially the same effect.

Re:Not really like VMWare

[Anonymous Coward]

Yes they may not have better performance over VMWare yet but once they make their emulator support multiple processors for the guest OS it will kick butt!

Excerpt from their TODO list posted on their website: Xen currently only supports uniprocessor guest OSes. We have designed the Xen interface with MP guests in mind, and plan to build an MP Linux guest in due course. Basically, an MP guest would consist of multiple scheduling domains (one per CPU) sharing a single memory protection domain. The only extra complexity for the Xen VM system is ensuring that when a page transitions from holding a page table or page directory to a write-able page, we must ensure that no other CPU still has the page in its TLB to ensure memory system integrity. One other issue for supporting MP guests is that we'll need some sort of CPU gang scheduler, which will require some research.
Link to the TODO

Re:Not really like VMWare

[borgheron]

This depends on what is meant by "port". Is it possible that they mean that they have written drivers to work on those operating systems to make it work?

I believe more information is needed on this before I make any decision as to whether it's better or worse than VMware.

GJC

Re:Not really like VMWare

[CTho9305]

Note also that they only benchmarked against VMWare 3.x. VMWare 4 is significantly faster than 3.

Re:Not really like VMWare

[ckaminski]

Right, and what no one seems to get out of this is that a lightweight kernel has been built that can do virtualization.

How much effort would it really take to put true virtualization back into this kernel, ala vmware?

Re:Not really like VMWare

[steve_l]

One thing you can do with OS porting for VMs is fix memory allocation so it is shared. under VMware, you tell the OS how much space it has, it thinks it owns it and it remains locked for the life of the app. So I have to allocate 700MB to Win2K+Visual studio to get our project's build to compile in a sensible time, leaving 300MB of ram for the hosting OS.

Whereas if they could co-operate, so that the OS asks the Xen kernel for memory, and releases it when needed, then they would all cooperate better.

Re:Not really like VMWare

[Anonymous Coward]

There are problems with VMWare as well. VMWare is only capable of of using certain network interfaces and other specialized devices that the HW may have. In addition to that, VMWare (3 and 4) dont always easily use the HW you want them to use. Try configuring one VM session to use one NIC and another VM session to user another NIC (it can be done but it becomes painfull). Everything has pros and cons and this is as good of a start as anything else I have seen. Who knows, maybe the changes for the 2.4.22 kernel can be permenantly embedded into the kernel source and we can simply active it when running under Xen so all future kernels can be run as well.

Re:Not really like VMWare

[swillden]

VMWare 4 is significantly faster than 3.

Unfortunately it's also significantly less functional than 3, at least on my hardware. 3 will do full-screen on my video cards and can use my serial ports, 4 cannot, and VMWare has been very unhelpful.

I paid for the upgrade to 4, but I actually run version 3 (I use it to run Windows for those few apps I don't have Linux equivalents for). I'm going to give them a few more months to see if they can fix the problems, and ask for my money back if they can't.

Re:Not really like VMWare

[mcrbids]

VMare can run your stock OS on a VM whithout the need to tweak it. [sic]

Not entirely correct. Conceptually, VMWare runs in its own virtual machine, so that the O/S does not know it's in a VM. Reality, however, is that the virtualization is not complete - so you can't just run any old goofy O/S - you have to use one of the supported systems, and which operating system you intend to run has to be explicitly specified when you set up a VM.

I can attest to this since I use VMWare W/S 3.x

Re:Not really like VMWare

[Wesley Felter]

To port an OS to Xen, you have to make some pretty low-level changes in the VM subsystem, etc. It's not just drivers.

Re:Not really like VMWare

[Torne]

We did benchmark against VMWare 4 (and ESX), but those versions have a clause in their licence agreements forbidding publishing of benchmarking results without VMWare's permission. They refused us permission, from which you can deduce who is faster. =)

Torne
(XenoXP porter)

Re:Not really like VMWare

[CTho9305]

Oh, cool :)

Re:Not really like VMWare

[borgheron]

Not good.

Re:Not really like VMWare

[Anonymous Coward]

I can't help you with the serial port issue, but I'm very curious about the fullscreen issue. I assume you have a Linux host? Do you have an SR# with VMware for the fullscreen problem?

post here, I'll check into it.

Re:Not really like VMWare

[swillden]

I can't help you with the serial port issue, but I'm very curious about the fullscreen issue.

I have an S3 Savage/MX video card (in my Thinkpad T21 laptop). What happens depends on the resolution I have Windows in the VM set to. If I have it at 1400x1050 (my screen resolution), the left-hand half of the screen is fine, but the right-hand half of the screen contains three copies of the right-most part of the left-hand half. Supposing the screen had the numbers 0 through 9 displayed left to right, in a font large enough to cover the screen horizontally, you would see something like:

0123454545

At lower resolutions, I don't see the duplication, but the display "vibrates", particularly along the right edge.

I assume you have a Linux host?

Yes. 2.4.20.

Do you have an SR# with VMware for the fullscreen problem?

I do, but it's on my laptop and I don't have access to it right now. I'll post it in the next couple of days.

As I mentioned, VMWare Workstation 3.2 has none of these issues.

Re:Not really like VMWare

[Anonymous Coward]

I'll look up the SR once you post it, but in the mean time:

Many X servers have various levels of bugs in the DGA2 extension. We just tell DGA2 what mode to use from the list it gives us. It's certainly possible that we caused the horizontal duplication, but I can't imagine how we could vibrate the display; your X server programs the video card, not us, and misprogramming the video card is the only way I can imagine to cause such a weird effect.

So why then, does it work in WS3.2 but not 4? I have a theory. One of the new features for WS4 is "depth emulation." The guest's depth does not need to match the hosts anymore. Can you make sure that your guest's depth is the same as the host's? I've heard second hand that there is a bug in certain drivers in certain XFree86 versions that only expresses if you choose a DGA2 mode that doesn't match your current non-DGA visual.

Lemme know how it goes.

Even if it doesn't fix the DGA issues, it'll make your VM faster.

Re:Not really like VMWare

[Anonymous Coward]

Did you find the SR#? Did changing guest depth/bpp to match host depth/bpp work?

I'd love to be able to track this down.

Re:Not really like VMWare

[Anonymous Coward]

BTW, if slashdot archives this story and stops allowing comments, you can reach me at:
kijiki0 (at) netscape (dot) net

Re:Not really like VMWare

[swillden]

Did you find the SR#? Did changing guest depth/bpp to match host depth/bpp work?

Sorry, I've been buried in work which required a usable VM, so I haven't had time to reinstall 4 in order to check the depth question. And it appears that I've lost my SR#s -- mainly because after exchanging fruitless e-mails for a few weeks, I gave up and dropped back to 3.2 so I could get some work done.

I should have time later this week to try the depth thing. If you'd prefer to e-mail rather than trying to monitor this /. thread, you can get me at shawn@willden.org.

Thanks for taking an interest!

Welcome

[show me]

I, for one, welcome our new Xen masters.

Re:Welcome

[beady]

I for one welcome our new Xen master welcoming overlords

Re:Welcome

[sjwt]

And I for one welcome VM that i dont have to pay a fortune for, i have yet to find a DOS emulator that works 100%, i wonder if theres even the slightest chance i can install DOS onto my current Xp1500+ machen, would be intresting
to say the lest.

Re:Welcome

[sharkey]

I'd like to remind them, that as a trusted TV personlity, I'd been useful in convincing people that it's OK to use a cut up beer can to fix excess play in the wheels of their expensive motorcyles.

Re:Welcome

[kzinti]

it's OK to use a cut up beer can to fix excess play in the wheels of their expensive motorcyles.

Handlebars. Pirsig used a piece of aluminum from a beer can to shim out some loose handlebars. Or at least he suggested it...

Re:Welcome

[sharkey]

Ah. It's been awhile since I read it.

Re:Welcome

[Glonoinha]

Try VMware. $300. www.vmware.com Free 30 day trial.
Lets you run DOS in a window on your Xp1500+ machine. Emulates Sound Blaster 16, I believe.

Honestly it kicks ass. If they dropped the price to $100 I imagine them selling them as fast as they could print them.

Re:Welcome

[kclittle]

In Russia we are all for Xen. (Comrade, it is not cold if vodka still pours...)

so its just an extra layer of abstraction?

[ikoleverhate]

So really, this is just an abstraction layer that means even the OS is unaware it's sharing hardware, so in theory theres no way for a malicious user to take advantage of other users. Pretty cool in a boring and limited sort of way. Kudos to the team who did it, I'm sure it's a real technological challenge. Not what the /. headline promised though ;)

Re:so its just an extra layer of abstraction?

[mrd_yaddayadda]

I'm inclined to agree. What a non-entity of both an article and an application.

I was all excited about some GPLed virtual machine host. Oh well.

Re:so its just an extra layer of abstraction?

[p3d0]

So really, this is just an abstraction layer that means even the OS is unaware it's sharing hardware...

No, that's not even close. The OS is very much aware that it has been ported to Xen. That's why this is different from VMWare.

Re:so its just an extra layer of abstraction?

[ikoleverhate]

The OS is aware it's been ported? Like as in "Hello dave. You want to port me to Xen? I can't let you do that Dave..." Changing the HAL (Hardware Abstraction Layer, and cunning pun) to be bound to Xen rather than actual hardware makes it different to VMware sure, but that doesn't imply any knowledge of the hardware, or more importantly, of the other OSs sharing the hardware. Like if you make a Java VM to work on a coffee machine's embedded system, are the user Java programs aware of coffee? No.

Re:so its just an extra layer of abstraction?

[p3d0]

Like if you make a Java VM to work on a coffee machine's embedded system, are the user Java programs aware of coffee? No.

Try to keep track of your own analogy here. If you port the VM to work on a coffee machine, the user programs are not aware of it; same with Xen. However, the VM is definitely aware of it, just like the OS is with Xen.

Re:so its just an extra layer of abstraction?

[ikoleverhate]

We're arguing semantics here, and I'm not sure if we even disagree... Read the following and decide. I reckon we're just arguing over the use of the word 'Aware'.

I don't think the ported sourcecode referring to Xen is enough to say the OS is 'aware' of it. You do.

I would say it was aware of it if it had something to compare it to ("this is a xen resource, this isn't") or there is some kind of choice ("ooh I'm being run in Xen mode, better do this").

Otherwise it *IS* just an abstaction layer as it doesn't know or use the information "I'm running on Xen". Only the programmer who did the porting is aware of that.

To return to my (admittedly badly conceived) analogy: The VM programmer knows he's ported it to the coffee machine. The VM only knows the inputs and output of the functions it uses to interface the hardware, whatever that hardware is and no matter that the function have been specificly coded for this VM and this coffee machine. It has no knowledge of what the hardware is.

I will admit I haven't studied the documentation on Xen to a deep enough level to be sure, but I would think that having the OS aware (and this time I'm using that to mean that the data is available, even to user programs. not just that it's in the source) that it is running on Xen would defeat the projects objective of ensuring the execution of one VM from affecting another. Spotting you were running on a Xen'd machine would be a big step towards breaking that.

Re:so its just an extra layer of abstraction?

[p3d0]

Perhaps it is semantics. I define "aware" with regards to Parnas' notion of information hiding. I think perhaps you are defining it in terms of runtime decision making performed by the code, but personally I don't find that to be a useful concept.

If I write some terminal display code that word-wraps at 80 characters, don't you agree that this implies the code is "aware" of the fact that my terminal is 80 columns wide? That knowledge is embodied in the code; given that code is not sentient, I can't think of a more useful concept of "awareness" than to say that a piece of code is "aware" of all the facts it embodies.

Likewise, if I must port an OS to Xen, thereby embodying in the code some assumptions about Xen, do you not agree that the code is therefore "aware" that it's running on Xen?

Re:so its just an extra layer of abstraction?

[ikoleverhate]

Well, at least we agree on what we disagree on ;) I would say that while the terminal code contains the 80 column info implicitly, that doesn't mean the it is aware of that information. I think theres is another useful definiton of awareness in this case - to be aware of something, the code has to have that knowledge available to itself for manipulation, and communication (eg to a user). I am not aware of my DNA sequence. It is implicit in my physical existance, but it is not information I am aware of. Layers of abstraction inherent in computer systems make this divide between implicit and explcit awareness of knowledge even more pronounced. Thanks for a nice intelligent discussion anyway, even if it was pointless ;)

Re:so its just an extra layer of abstraction?

[p3d0]

Ok, maybe your definition of "awareness" is useful, but I'm not going to start using it any time soon. :-)

Take care.

Re:so its just an extra layer of abstraction?

[ikoleverhate]

hehe, understood ;)

multiple desktops

[Anonymous Coward]

The reason i want multiple desktop screens is because just having one is not efficient when i have to work on multiple different things because they need to get done on time. This will save me from having no time left for not doing quality work when I am not at home.

Re:I bet it's not Open Source...

[Aspasia13]

From the page (which hasn't bet /.'ed... yet)" you can read that Xen itself is GPLed:

"Modern computers are sufficiently powerful to use virtualization to present the illusion of many smaller virtual machines (VMs), each running a separate operating system instance. Successful partitioning of a machine to support the concurrent execution of multiple operating systems poses several challenges. Firstly, virtual machines must be isolated from one another: it is not acceptable for the execution of one to adversely affect the performance of another. This is particularly true when virtual machines are owned by mutually untrusting users. Secondly, it is necessary to support a variety of different operating systems to accommodate the heterogeneity of popular applications. Thirdly, the performance overhead introduced by virtualization should be small.

Xen is a virtual machine monitor for x86 that supports execution of multiple guest operating systems with unprecedented levels of performance and resource isolation. Xen is Open Source software, released under the terms of the GNU General Public License. We have a fully functional port of Linux 2.4 running over Xen, and regularly use it for running demanding applications like MySQL, Apache and PostgreSQL. Any Linux distribution should run unmodified over the ported OS.

With assistance from Microsoft Research, we have a port of Windows XP to Xen nearly complete, and are planning a FreeBSD 4.8 port in the near future (volunteers welcome!). "

ob. Beowulf cluster comment

[Vic Metcalfe]

Finally I can create a Beowulf cluster without the clutter of all those machines!

Re:ob. Beowulf cluster comment

[ERJ]

I know that the parent was meant to be funny, but there is some truth behind it. This could make work testing / debugging cluster computing software a heck of a lot easier. Obviously the performance would not be there, but the ability to simulate multiple computers without needing access all the time to the multiple computers could be helpful.

Then again, you could probably just run several copies of the same process in one environment, but it would be nice to be able to simulate the seperate user space that multiple computers has.

Par For The Course

[Anonymous Coward]

It's not too often that the article delivers on what the /. headline promises.

The only thing I'm wondering...

[metroid composite]

...is if you can copy and paste between the OSs. That always annoyed me when I had Linux open remotely through a Windows machine. I had to leave a submission form open on my website as a "back door" to copy stuff in. This goes for things like Gnumeric to Excel data too.

Re:The only thing I'm wondering...

[Anonymous Coward]

I regularly copy and paste into and out of Putty and VNC. Though I don't know if you can do it with Cygwin, but I don't have much need for it.

Also check out Synergy if you got multiple machines each with its own monitor, and one set of keyboard/mouse. I haven't used it, but its supposedly to allow you move your mouse past the edge of the screen to switch between machines, and also shares the copy/paste buffer between machines.

Re:The only thing I'm wondering...

[Andrewkov]

You can cut and paste from the Windows Telnet program (right-click on the title bar), or are you using an X display?

Re:The only thing I'm wondering...

[nacturation]

Clearly this won't be possible. The operating system is running as if it were on a separate machine, so it would be like trying to copy and paste from one isolated PC to another by pressing Ctrl+C on keyboard A and then hitting Ctrl+V on keyboard B. Aint gonna happen.

Clipboard operations are done within the operating system itself, so unless you have some kind of clipboard daemon running which other OSes can connect to and poll your clipboard contents, you're SOL on this one. Stick to something like VNC, ssh, telnet, etc. which allow for this kind of thing.

Re:The only thing I'm wondering...

[0x0d0a]

It will require the addition of high-level software. It's not possible with just a VM system.

Re:The only thing I'm wondering...

[caseih]

Since xen is really a machine monitor, not an operating system, you won't be able to simple cut and paste between them. However, there are a number of utilities out there that can link clipboards (windows and linux) over a network connection. One example is MpCb, at http://www.idata.sk/~robo/mpcb/.

With software like MpCb, and if Xen supports virtual terminals, you could easily switch back and forth between, say XP and Linux. I know our CS computer labs would love that.

Depending on how flexible Xen is, one could run two video cards, keyboards, and monitors, and have completely separate terminals and OS's running on e one machine. This is seriously cool.

Re:The only thing I'm wondering...

[gl4ss]

Theres free software for that, called synergy. Lets you use one kb/mouse for multiple computers(running x or mswindows maybe other systems too, though no beos port: ) and lets you cutnpaste text as well.

Re:I bet it's not Open Source...

[sjwt]

If you had botherd to open the Xen link,
and read the page which was a whopping
8100 bytes in size you would of noticed
"Xen is Open Source software, released under the terms of the GNU General Public License. We have a fully functional port of Linux 2.4 running over Xen, and regularly use it for running demanding applications like MySQL, Apache and PostgreSQL. Any Linux distribution should run unmodified over the ported OS."

rather then just trolling off the two words "Microsoft Research"

Re:Paid Advertisment

[Anonymous Coward]

It's GPL software.

But the real question is...

[tie_guy_matt]

How do you pronounce Xen? Is it "Zen" like xenon, or "X-en" like x-ray?

Re:But the real question is...

[Anonymous Coward]

I'd imagine the x would be pronounced the same as it is in the name Xavier. A sort of sped up "ks" sound.

Re:But the real question is...

[Dougie]

I would have thought it would have been pronounced like Zen, as in Xine, or a "real world word" xylophone http://dictionary.reference.com/help/ahd4/pronkey. html

Any thing along the lines of an X as in galaxy is just too much effort to pronounce.

Doug

Re:But the real question is...

[Anonymous Coward]

It's probably pronounced as "zen", not as "zeen" (which is the way xenon is) or "ecksen".

Re:But the real question is...

[Anonymous Coward]

It's "TEN" as in OSX

DOWNLOAD Here

[Anonymous Coward]

http://sourceforge.net/project/showfiles.php?group _id=86024

http://prdownloads.sourceforge.net/xen/xendemo-1 .0 .iso?download

http://www.google.com/search?q=xendemo-1.0.iso

Comparison with other free VMs?

[Laur]

So it looks like this is the third (or fourth) free VM for Linux, the others being Plex86 (and a different fork here) and User Mode Linux. Does anyone have a good comparison of these three? I know Zen compared UML on their site but not plex86. I'm not really sure of the differences between them, particularly the different versions of plex86 and UML (Zen explained their virtualization process pretty well on their site). Which is the best choice for different scenarios? It looks like Zen is the winner for running Linux as the guest OS, and the original Plex86 (first link) is the only one which offers a free choice in guest OS's.

Re:Comparison with other free VMs?

[emj]

Well:
User Mode Linux

1. Is linux Ported to the linux syscalls
   
2. Has been ported partially to windows as well
   
3. No hardware access only software modificatade to be hardware
   
4. Is just Linux
   
5. Extremly Easy do install 2 Files to download and then start NO configuration
   
6. Very Fast emulation of linux, 10% Slow down while compiling the kernel
   
7. Support SWAP partition as memory for virtual machines
   

Plex86

1. Is X86 & Bios ported to C
   
2. Runs multiple OSs
   
3. Runs on multiple OSs
   
4. Very hard to conigure
   
5. SLOW!
   
6. Thourgh Emulation of the hardware no just software
   

Xen

I have not run this Emulator but it looks like it is more or less an OS (it has to have support for Graphics card/NICS/MMU to virtualize them). So it seems like it work in the same way as UML but tries to make the big change in the operating system running bellow instead ontop. Now I need to read that paper more before I can say anything, but they do alot less changes to the kernel than UML (~2700 lines of codes in the kernel).

Re:Comparison with other free VMs?

[Anonymous Coward]

Don't spell it the way you pronounce it -- spell it the way it's written. Xen, not Zen. Sheesh. What's next? Writing the possessive "OS's" instead of the plural "OSes"?

Re:Comparison with other free VMs?

[Laur]

Thanks for the info, however you seem to be confusing Plex86 with Bochs. They share a lot of code, however while Bochs emulates the entire hardware (and thus can be used on any platform) Plex86 virtualizes many things and so can only be run on x86 hardware, but is supposed to have much better performance because of this. Also, AFAIK Plex86 is only available for Linux (as the host OS). The original Plex86 should run any x86 OS. The new Plex86 fork simplifies the virtualization even more and only runs Linux, supposedly with an even greater performance increase, providing basically the same thing as UML but in a different manner.

What I was interested in was performance comparisons between the different Plex86 versions, UML, and Xen, as well as some info on how each goes about virtualizing the hardware (which is better and why). I was playing with Bochs just a few weeks ago (on a 600 MHz ibook, Linux as the host, Windows 98SE as the guest) and found it unusably slow. I haven't played with Plex86, UML or Xen yet, but I plan to as soon as I get the free time. ;)

Re:Comparison with other free VMs?

[emj]

Now that I have tried it out:
Xen: fastest
UML: fast
bochs: slow

I have ran each on 300MHz and bochs on a 2GHz..

Re:Comparison with other free VMs?

[Laur]

What about Plex86? Bochs is very slow because it is emulating every tiny bit of the hardware, including all I/O devices. This isn't neccessary if you are just running x86 on x86, Plex86 streamlines it quite a bit and virtualizes many things. The newest Plex86, just for running Linux on Linux, is supposed to be very lightweight.

Re:Comparison with other free VMs?

[Asmodeus]

Note that there is a fork of Boch which uses some of the Plex86 code to speed it up. This should potentially give the best of both worlds, but doesn't appear to be well documented or readily available; read the Bochs web site/email archive for details. BTW, Xen appears to be identical in philosophy to the new Plex86 fork, just Xen have the access to port WinXP as well.

Virtual Virtual Machine?

[Stiletto]

So when do I get a Virtual Virtual Machine, to allow me to run all these VM's at the same time? Because as we all know, it makes much more sense to do this, than, oh I don't know... Buy another $299 computer?

Re:Virtual Virtual Machine?

[The_ForeignEye]

True, for home use you would probably be better off by just getting another cheap PC and switch between them (for as long as you don't mind having two computes chewing up power, space, and noise).

However, in a corporate environment, virtual machines make a lot of sense. You can purchase a beefy server that gives you all the redundancy and fail-over features that you need to run business critical applications, and at the same time you can "divide" that machine into virtual servers, each one performing a different function, and benefitting from the redundancy and fail-over. Really cheap and cost effective solution!!

Re:Virtual Virtual Machine?

[Glonoinha]

You know that is funny, I was thinking that the other day ... can you cascade VMware? I mean on a VM running Linux on a Windows 2000 host, can you install VMware for Linux and create a VM and install Windows 2000 on it?

Or would that be the equivalent of putting a portable hole into a bag of holding (or vice versa,) causing either a dimensional rift which either imploded sucking everyone and thing in immediate area into another plane or exploding and destroying everything within immediate area.

Re:Virtual Virtual Machine?

[Ceyan]

You're right it does. Especially when you consider that $299 computer takes up extra space, and you've got to spend the money on the KVM (or another monitor/keyboard/mouse). Plus why a machine set up specifically to be a honeypot that you only need once a month when you can use a VM? I've got about 30 VMs of machines of various OS types at various patch levels with various software on each VM. Imagine how much space that would take up using individual PCs.

Re:I bet it's not Open Source...

[Anonymous Coward]

What's the bet the Xen version of XP will have changes to product activation to ensure you can't just clone multiple copies of an activated installation (identical hardware remember). Then again I can't really think how they could prevent it on current hardware, but I'm sure they try and charge for each copy running under Xen.

Nice jab at Mozilla!

[seanmeister]

"....and regularly use it for running demanding applications like Apache, PostgreSQL and Mozilla."

That's kinda funny, lumping a web browser in the same category as server apps designed to handle gazillions of users.

Come on, it's not THAT bloated!

Re:Nice jab at Mozilla!

[bobv-pillars-net]

Come on, it's not THAT bloated!

I've run Apache and PostgreSQL on a desktop machine, but I wouldn't dare run Mozilla on a server.

Re:Nice jab at Mozilla!

[bobv-pillars-net]

(cringe)

Okay, I've been trolled. After RTFA, I realize that it didn't even mention Mozilla.

Re:Nice jab at Mozilla!

[seanmeister]

No, you haven't been trolled. Not by me, anyway. The text I quoted appears in the article as posted on /. ... I R'd TFA a bit later and saw that the actual article didn't mention Moz!

Re:Nice jab at Mozilla!

[Chester K]

That's kinda funny, lumping a web browser in the same category as server apps designed to handle gazillions of users. Come on, it's not THAT bloated!

At least they didn't mention emacs.

Re:Nice jab at Mozilla!

[hurtta]

"....and regularly use it for running demanding applications like Apache, PostgreSQL and Mozilla."

That's kinda funny, lumping a web browser in the same category as server apps designed to handle gazillions of users.

Come on, it's not THAT bloated!

Well, actual text is:

We have a fully functional port of Linux 2.4 running over Xen, and regularly use it for running demanding applications like MySQL, Apache and PostgreSQL. Any Linux distribution should run unmodified over the ported OS.

Mozilla is not mentioned :-)

Re:Nice jab at Mozilla!

[evilviper]

That's kinda funny, lumping a web browser in the same category as server apps designed to handle gazillions of users.

Come on, it's not THAT bloated!

I'm sorry to say that it is indeed that bloated... I had also forgotten how bloated it was, until recently a system died, and I had to move everything over to a 233MHz temporarily.

Just about everything still runs just fine... Abiword pops up quite fast, gaim isn't a problem, ssh seems just as fast as ever, and even GTK-Gnutella doesn't seem any less responsive... Then there is mozilla...

Mozilla, so bloated that it takes 5 minutes to load. Mozilla, so slow it frezes for 30 seconds as it tries to open a new webpage. Mozilla, so unresponsive that you might as well take a coffee break after clicking on an e-mail message.

Sure, I know how benign Mozilla can seem when it's the only CPU-intensive process running on a multi-GHz machine, but loose some system memory, and drop the clock speed, and you'll very quickly see what a hog it really is.

Mozilla, I think I'll slam my head against the wall just for something to do to kill some time after I click sumbit...

Very cool indeed

[bigjnsa500]

Very cool, but why is Microsoft involved? Getting their hands dirty with GPL?

"Cats and dogs sleeping together, MASS HYSTARIA!"

Re:Very cool indeed

[Rob Riggs]

Very cool, but why is Microsoft involved?

I would guess it is because Microsoft has a modest edge over Linux International when it comes to funding research grants.

Re:Very cool indeed

[Anonymous Coward]

why is Microsoft involved?

Because you're a moron if you think in reality microsoft actually give a crap about GPL and linux being a "problem".

The only reason microsoft care about the GPL is because there's a whole bunch of nifty code in it that they can't steal and manipulate, like they usually do - and that, I am sure pisses them off!

It is possible to live life without the Tin Foil hat ya know!

I hate Xen

[nnnneedles]

It was all good until you came to Xen, I mean those levels just sucked..

Re:I hate Xen

[lacerus]

I even stopped playing for over a year because I didn't figure out how to jump from one of these stupid flying islands to the next!

Denali?

[spotter]

How is Xen (published this year at SOSP) different than Denali (published last year at OSDI, OSDI and SOSP are basically equivalent level conferences, held in alternating years)

They both seem to be scalable hardware monitor type virtualization architectures, in skimming through the paper I was left wondering what makes Xen special.

Re:Denali?

[Anonymous Coward]

Xen runs real applications, unmodified, on a real OS.

Re:Denali?

[Deagol]

They specifically compare the 2 in the paper they link to.

Re:Denali?

[spotter]

ok, good point, I didn't buy the argument at first of

"Denali does not fully support x86 segmentation although it is exported (and widely used1) in the ABIs of NetBSD, Linux, and Windows XP."

as I was like "Xen has to port linux to Xen, just as Denali has to port linux to Denali", but I realize your point of running existing x86 binaries on top of Xen. Strange that I would ignore/miss that, as that's a similar point that I try to make in my own research.

Re:Denali?

[spotter]

as pointed out in reply to other post, I didn't initally buy the initial argument at first. In rereading it more carefully, it makes sense.

Not sure why I didnt realize that b4, probably didnt focus enough on the paper as was busy with submission for NSDI, and came in with the "this can't really be different from denali, can it" bias.

the libOS argument: I'm iffy on this argument. Denali's OSDI paper "claims" there's a linux port in progress. i.e. hand-wavey so I can see why one can make the claim.

the paging argument: I'm not sure I buy the argument at all. i.e. if one views the VMM as an OS (at it is in many ways) what's the big difference in it doing the paging vs. the gues OS's. but my lack of knowledge about paging might mean I'm speaking out of my butt.

namespace argument: I don't like this argument.

first they claim in the arugment that

"In contrast, we believe that secure access control within the hypervisor is sufficient to ensure protection; furthermore, as discussed previously, there are strong correctness
and performance arguments for making physical resources directly visible to guest OSes."

only refence to this argument that I see is this small point, which they don't seem to back up anywhere (though that's just from a quick search through the paper for "correct")

"Even on cooperative machine architectures, completely hiding the effects of resource virtualization from guest OSes risks both correctness and performance."

I don't buy the argument because from a security pov, if everything is done via a virtualized call, its simpler to implement and simpler to understand (if it's not named, you can't find it, hence can't touch it). Similarly, the code is generally simpler (simple hash table lookup replacing original location with new location).

On the other hand, performance can possibly be an issue, my experience with related work in OS's itself (virtualizing resources for security purposes) is that the overhead is minimal. However, I can deffinitly imagine that a VMM is different as it constantly has to do the transalation, while the OS doesn't.

I could be wrong on my take on this argument, but I didn't see much in the paper supporting it. (again, just skimmed through, will try and read more carefully later)

anyways, just my 2 cents. Point 1 seems to be a good enough point overall anyways.

Re:Denali?

[Anonymous Coward]

There are several similarities and differences between Xen and Denali. The Xen paper's comparison is fair, but not extensive. Some more comparision points are:

• Denali was the first to push the idea of paravirtualization, namely changing the virtual architecture to gain speed/scalability advantages, but necessitating porting the guest OS (i.e., rewriting the arch/ directory of linux). Denali came out in OSDI 2002. Xen uses the same idea, albeit with different virtual architecture changes. Xen will be presented in SOSP 2003 in late October.
• Xen was the first to convincingly port a real guest OS to a paravirtualized VMM and demonstrate great performance. Denali has since done the same, running NetBSD. Xen's working on more guest OSs than Denali.
• Both Xen and Denali are type 1 VMMs, in that the virtual machine monitor runs directly on the hardware. VMware ESX is also a type 1, but VMware workstation and UML are closer to a type 2, since they run with the support of a host OS.
• Xen supports unmodified x86 guest OS binaries. Denali requires recompilation of guest OS binaries, but doesn't require any source code changes.
• Xen focuses on precise resource accounting. Denali focuses on scalability, supporting a very large number of concurrent VMs.
• The two projects have gone in different directions since these papers, from what I understand...stay tuned for more papers, I guess!

Re:Denali?

[p3d0]

Good grief, brother, they have devited almost the entire section 2 of the paper to a comparison between Xen and Denali. If you missed that, I have grave doubts that you actually "skimmed the paper".

Re:Denali?

[spotter]

argh, read my follow up posts to the other comments along the same lines. maybe I'm a bit too picky.

wow this is too cool

[Anonymous Coward]

wow this is too cool!

First Post!!!!

Any VMWare user comments?

Re:morons tout creator's 3d+ newclear power plan

[Anonymous Coward]

What is this, a terrorist pseudorandom code thing? Release NT/IE virus if the DOW is up today? The misspellings alone have enough potential double meaning to get across a whole lot of info...

For all those not just out of nappies/diapers...

[jantheman]

Oh, you mean this then?

(sry NYI)

When do we get a Xen compiler?

[James McP]

Okay this is neat for the mass roll-outs but for anyone doing R&D or otherwise living on the bleeding edge they're going to need a compiler that handles "make -xen" to generate those new kernels.

I could see this being he11a kewl with the kernel-based web servers for maximum security sandboxing.

Yeah, whatever...

[amarodeeps]

Xen lets you run multiple operating system images at the same time on the same PC hardware, with unprecedented levels of performance and resource isolation.

Ha. That's easy for you to say. Every time I try to run in Xen I have to dodge Headcrabs and be on the look out for Vortigaunts, not to mention the dreaded Gonarch.

User Mode Linux

[shird]

Sounds an awful lot like usermode linux. Get a vesion of the kernel/OS which makes calls to a host virtual machine rather than directly to the hardware for privileged instructions. Everything else can run directly through the CPU without being emulated because it is running in 'usermode' (or ring 1).

Re:User Mode Linux

[kervel]

except that you don't run it on linux but on some layer directly above the hardware (user mode linux is linux on linux). that explains the better performance too

BitTorrent link available

[stab]

Grab the BitTorrent from here (and leave your windows open for a while!)

Why Microsoft+Intel? NGSCB backward compatibility!

[NZheretic]

A quick glance over Xen group's paper leaves me very impressed with the performance these techniques can achieve. That the Xen group has decided to relase the code under the GPL leaves me very greatful. However, that both Intel Research and Microsoft Research has funded it, leave me somewhat concerned.

As I have stated before about Microsoft's purchase of Connectix's Virtual Server technology

In my opinion Microsoft's acquisition of Connectix's Virtual Server technology has very little to do with running any other vendors operating system.

Microsoft needs a Virtual Server for backward compatibility for it's NGSCB ( Next Generation Secure Computing Base ) DRM ( Denial of Rights Mechanism ) platform.

Just as Microsoft's XP backward Win9x compatability opens up many locally exploitable API to gain SystemLocal privilege access, to the point where many programs need Adminstrator privilege to run, existing XP and win2k software would open up too many opportunities for helpfull hacker to bypass Microsoft's NGSCB DRM mechanisms.

Microsofts all too obvious solution is to provide a "Virtual" PC mode, running a modified XP and WinME, with the NGSCB providing virtual filesystems and hardware access. All, access of course, with the NGSCB DRM scanning and control.

Where do you want to go tomorrow?

The Electronic Frontier Foundation are about to publish a paper criticizing a component of the "trusted computing" technology promoted by Microsoft, IBM and other technology companies, calling the feature a threat to computer users..

Hardware support

[Anonymous Coward]

According to the README, it requires special hardware drivers and is not targetted at desktops. Don't expect stellar graphics performance. VMWare *does* give you something for the money.

Hardware support
================

Xen is intended to be run on server-class machines, and the current
list of supported hardware very much reflects this, avoiding the need
for us to write drivers for "legacy" hardware. It is likely that some
desktop chipsets will fail to work properly with the default Xen
configuration: specifying 'noacpi' or 'ignorebiostables' when booting
Xen may help in these cases.

Xen requires a "P6" or newer processor (e.g. Pentium Pro, Celeron,
Pentium II, Pentium III, Pentium IV, Xeon, AMD Athlon, AMD Duron).
Multiprocessor machines are supported, and we also have basic support
for HyperThreading (SMT), although this remains a topic for ongoing
research. We're also looking at an AMD x86_64 port (though it should
run on Opterons in 32-bit mode just fine).

Xen can currently use up to 4GB of memory. It's possible for x86
machines to address more than that (64GB), but it requires using a
different page table format (3-level rather than 2-level) that we
currently don't support. Adding 3-level PAE support wouldn't be
difficult, but we'd also need to add support to all the guest
OSs. Volunteers welcome!

We currently support a relatively modern set of network cards: Intel
e1000, Broadcom BCM 57xx (tg3), 3COM 3c905 (3c59x). Adding support for
other NICs that support hardware DMA scatter/gather from half-word
aligned addresses is relatively straightforward, by porting the
equivalent Linux driver. Drivers for a number of other older cards
have recently been added [pcnet32, e100, tulip], but these are not
recommended since they require extra packet copies.

Re:Hardware support

[Torne]

Actually, the hardware drivers are not very special; the device core is taken from Linux as much as possible (gotta love the GPL) so the drivers for SCSI/IDE/ethernet are slightly modified Linux drivers, nothing more. One of my colleagues ported ten or so network drivers from Linux to Xen in a day. =)

Torne
(XenoXP porter)

Microsoft Research

[non]

in case anyone forgot, not only is microsoft research their neighbor, but it was also the first microsoft research center outside the us. wired has more about what you can get for $80 million.

Re:I bet it's not Open Source...

[samj]

I'll bet even the binaries won't see the light of day, so it's all good and well to say it works, but that's not going to help us any time soon.

xp licence/activation .

[leuk_he]

If they ever release the XP binary you will have to bet that it will be some hight priced version of xp that runs it. Or else you will run into all kind of license problems.

If it will be a concurrent of The virtual machine solution they bought from connectix This will not be released.

And if it will work with a standard XP home/pro you will have all kind of activation loopholes like in vmware.

Re:xp licence/activation .

[Elwood P Dowd]

If Xen were BSD Licensed, then Microsoft might have just rolled it into Windows XP and their virtualization software. Probably only if they could find a way to make it so that Linux wouldn't run on it.

However, it's GPLed, so I can't imagine it'll ever see common use as a host for Windows. Unless they give MS a special dispensation.

Source code not available

[avenj]

The source code is only available via BitKeeper, as far as I can tell from their site. The BitKeeper tools are not free software and cannot be used by anyone who has contributed to a competing product, according to the license for the free-as-in-beer version of BitKeeper. Is there somewhere else to get source and I'm just missing it?

Re:Source code not available

[greygent]

Whiner.

Re:Source code not available

[dmaxwell]

If he is correct about the license then it is a legitimate concern. If someone contributes to say Subversion, then they are locked out of contributing to unrelated projects that are managed with Bitkeeper unless the maintainers have an alternate method for accessing the development tree.

Re:Source code not available

[penguin7of9]

Worse, if someone agrees to the Bitkeeper license and then contributes to Subversion then Subversion may be in trouble.

Re:Source code not available

[Anonymous Coward]

No, if someone agrees to the Bitkeeper license and then contributes to Subversion then they are in trouble.

Re:Source code not available

[Anonymous Coward]

You are such a whining little monkey it's unbelievable. Please go home and play with your D&D figurines instead of bothering us here.

Re:Source code not available

[Anonymous Coward]

They may be in trouble, too. But that wouldn't get Subversion into the clear: if Subversion is contaminated by proprietary source code, then Subversion is in trouble.

Re:I bet it's not Open Source...

[mummers]

Correct. The 'Gates Public Licence' (TM). Well known to the DoJ and others.

Mod -1: Flamebait

Not very useful for Linux fans like myself :(

[urdak]

As a person who prefers Linux over Microsoft Windows, and works in Linux almost all the time, how will I be able to use Xen? I don't want to run Linux inside Windows - I want to run Windows programs inside Linux! And since Microsoft is not likely to sell the ported XP anytime soon (why would they??) I probably won't be able to do that.

It sounds like Xen will only be of use to Microsoft users who want to try out Linux... That's good, but there are already plenty of solutions for that (e.g., Knopix) and they are not really useful to me.

It's also not clear how, if at all, Xen can support displaying graphics from Windows (say, a Microsoft Word window) inside X-Windows.

Re:Not very useful for Linux fans like myself :(

[arkane1234]

It sounds like Xen will only be of use to Microsoft users who want to try out Linux... That's good, but there are already plenty of solutions for that (e.g., Knopix) and they are not really useful to me.

It also sounds to me like you didn't even remotely read any of the website or article.

D'oh

[stratjakt]

I was all excited till I got the part about it requiring kernel modifications.

So now, as I understand it, this can really only run multiple instances of linux, or perhaps BSD.

I was hoping for something that can run Windows and OS/2, BeOS, side-by-side with these super-fantastic performance levels.

Now, if you require their code in the kernel for it to work, would that mean MS would have to GPL the NT kernel to make it compatible? (fat f'ing chance)

Would the performance stay where it is if OS's other than linux 2.4.22 were used? Or, in other words, how much of the performance is due to linux-specific tweaks?

How well does something like this share a network card? Could I run multiple servers (isolated from one another) on the same box? How can I play with this at home?

Oh well, good job anyhow. Kudos on getting mozilla running under it, too. I can barely get that to run on a machine with a single system image.

That's not what he was talking about...

[Anonymous Coward]

The actual execution of Product Activation wouldn't be an issue. The business case of how to handle product activation, and how to go about distributing a modified XP Binary plus activation authorization.

Re:That's not what he was talking about...

[Anonymous Coward]

I see what you mean, that's a very good point.

After all, Microsoft have already been burnt by the volume licensing issue with the onset of key generators and no activation required. I can see that they'd be reluctant to repeat that mistake again with this.

Intersting

[Anonymous Coward]

How much additional work would it be to enable the OS to run without porting, and sacrifice some of the performance?

Seriously, why do I want Linux AND OS X?

[caveat]

Really, besides moral arguments about "freedom", why on earth would we want to migrate users en masse, or even at all, from OS X to linux? I used LinuxPPC/YDL for ages on my 6400, and I don't buy the arguments about linuxPPC being a more stable or more mature OS than X - sure, the UI and other high-level stuff might be a little rougher around the edges, but Darwin, IMHO, is already better than LinuxPPC in every arena, usability, robustness, hardware support, ease of configuration, etc etc etc (then again, I'd also say BSD is better than Linux...*hnads over the saltshaker*).
Software? Seems pretty much every OSS package that was ready for linuxPPC has an OS X port, even more so with the X11 release...really, X is already a better deskop *nix than linux, and I don't think I, at least, will ever be convinced otherwise.

Re:Seriously, why do I want Linux AND OS X?

[Anonymous Coward]

Really, besides moral arguments about "freedom", why on earth would we want to migrate users en masse, or even at all, from OS X to linux?

Agree 100%, we are talking about moving from Unix to Unix. It's only a question of personal preference as I said to the other person.

I used LinuxPPC/YDL for ages on my 6400, and I don't buy the arguments about linuxPPC being a more stable or more mature OS than X - sure, the UI and other high-level stuff might be a little rougher around the edges, but Darwin, IMHO, is already better than LinuxPPC in every arena, usability, robustness, hardware support, ease of configuration, etc etc etc

I think it is difficult to say which one is better in stability etc. I am using both but since none crashes, I cannot really tell. Also, concerning hardware support, you need to take into consideration that Apple (same thing applies to SGI, Sun etc.) have a limited hardware to support. On the other hand, Windows and GNU/Linux do not provide hardware and have more variety to support. So, although I agree that OS X has better hardware support, taking the last stmt into consideration and that Apple has the hardware specs, GNU/Linux has quite good hardware support for what it has access to.

Software? Seems pretty much every OSS package that was ready for linuxPPC has an OS X port, even more so with the X11 release...

Agree 100%. From a user (not administrator/developer point of view) OSS runs on OS X.

Re:Seriously, why do I want Linux AND OS X?

[Anonymous Coward]

> From a user (not administrator/developer point of view) OSS runs on OS X.

What I meant, is from a user point of view, the underlying OS (GNU/Linux, *BSD, Mac OS X, etc.) does not matter.

Web Mirror and torrent link

[stab]

The university's link is a bit hosed, so I've mirrored the site and paper here, and you can grab the ISO via BitTorrent from here

Re:Web Mirror and torrent link

[patbob]

SO via BitTorrent from here

Thanks for posting the bittorrent link. I'm beginning to hate hosing sites, especially for a large download like this.

Re:Web Mirror and torrent link

[ChrisUK]

> Thanks for posting the bittorrent link. I'm beginning to hate hosing sites, especially for a large download like this.

Actually, it wasn't you; most of that area of Cambridge suffered a prolonged power cut today. At the time of the posting, the Computer Lab had connectivity from 1 (one) backup DSL line, and nothing else.

Re:Web Mirror and torrent link

[patbob]

Actually, it wasn't you

Well, I didn't think it was me alone :-). I kinda figured I and a few hundred thousand of my slashdot compatriots might have collectively done 'em in.

Come to think of it.. didn't someone express some concern that a factor in the US east coast blackout was some virus making computers suck more power than normal.. I wonder.. :-)

Re:I bet it's not Open Source...

[larry bagina]

The GNU General Public License is NOT Open Source.

someone

[Anonymous Coward]

needs to port this to Windows. It would be great for Windows users to see this sort of thing (and compare performance versus Bochs and VMWare.)

New video card for gamers?

[GoofyBoy]

From the title I thought it was about a new video card which was built specifically for Half-Life 1/2.

Know what?

[Anonymous Coward]

I'm with you 99%.

The irony...

[Alkarismi]

From the site:

"Work on Xen is supported by UK EPSRC grant GR/S01894, Intel Research, and Microsoft Research.

Am I dreaming ...

[gonvaled]

or is MS actually collaborating in a GPLed product?

Have we actually reached the inflexion point were stupid confrontation will be slowly replaced by fruitfull collaboration?

When is MS going to release their own GNU/Linux distribution?

Re:I bet it's not Open Source...

[frp001]

I'll bite... As "Not" stated here...

This means that VMWare goes out the window

[Jjaks]

I use VMWare at work because I still must use Windows for some dreary office tasks. If this works well, I would be able to chuck VMWare out! That is a very good thing, because VMWare is expensive and performance is pretty lackluster in my opinion.

Re:This means that VMWare goes out the window

[Anonymous Coward]

What are you talking about?! Performance under VMWare is excellent, I think you forget it's a goddamn PC emulator (and a damn good one) ... I do agree it's a bit expensive, it used to be $100 bucks - I think if they kept the personal license at this level, they would actually have more sales ...

Re:This means that VMWare goes out the window

[base3]

I agree with that--I bought personal licenses for both Linux and Win32 when they had the good pricing. But when it became as much as I paid just to upgrade, I stuck with the old versions.

No.

[Anonymous Coward]

First of all, Xen doesn't run ON windows.. its an OS itself. Secondly, the article specifically says that windows is being ported to Xen.

Wow

[Anonymous Coward]

Microsoft employs whole UNIVERSITIES! Now that's pretty amazing. Also scary. Makes me think of them as a monopoly even more now...

Even if they're nice to some people, doesn't mean they aren't a monopoly and aren't mean to some people...

Benchmarks?

[LinuxInDallas]

Is it just me or is the use of SPEC INT2000 a little odd? They claim they use t because it is the most CPU instensive as opposed to other benchmarks that are more IO intensive. Real apps are a mix of both. Wouldn't you want to include IO performance?

How useful can it be...

[LilMikey]

if it requires special kernel compiles? That's fine for Linux but, assuming Xen proves useful at all, how much future support will it have from MS? I'm surprised they helped out at all in light of their recent purchase of VirtualPC.

In the near term though, it'll be pretty cool to run XP and Linux... someday as "for the moment, you only get to run multiple copies of Linux on Xen"

that's the point!

[RMH101]

hate to point out the obvious, but the whole point of FAIB is that:
a) if you want to, you can brew your own - it's open source, kinda
b) if you can't be bothered or don't have the mad skills, you can buy it for cheap at the local shop
c) if you want, you can choose to drink expensive imported stuff because you like the taste

"embrace and ..."

[s4ms4r4]

maybe M$ intends to bundle an "approved" Linux distribution at some point?

Hold up a minute...

[Anonymous Coward]

Just lemme get my Crowbar....

my KVM switch

[RMH101]

lets me do exactly this, not quite as portable though but i trust the HAL better this way...

Modified Kernels? No Thanks

[nurb432]

The idea sounds frightening.. having to modify your kernels just to run it as a VM..

And they are porting XP? how does one even get a hold of a modified kernel. is that even legal to distribute to another person that doesn't have the same agreement with Microsoft..

Re:Modified Kernels? No Thanks

[patbob]

The idea sounds frightening.. having to modify your kernels just to run it as a VM

This shouldn't be too frightening, after all, most payroll systems run on IBM OSs running on top of a VM layer, and the OS kernel was modified to play well with that VM layer many years ago. It can be both very reliable AND reasonably performant.

Really, the only other way to run virtual machines is to emulate the hardware. Making the OS be a client of a lower level resource services layer is a way better solution that trying to emulate the idiosyncricies of hardware just to avoid a refactoring of the kernel functionality.

If these folks have speced the resource layer out correctly, it may be possible to implement the kernel with the interface built in.. sort of like MS's HAL (oops.. using MS as an example.. was that a bad thing? :-), and an own-the-harware default resource layer implementation. With a little luck, they even thought ahead and abstracted the device driver interface to their hypervisor, allowing new devices to be easily integrated into it.

Of course, doing this does sort of commoditize the OS kernel and kernel device drivers, which probably won't sit too well with some.

hmmmm...

[NickFortune]

I can see it now...

The Xen layer wil have some sort of deep down DRM. It'll be set up so XP won't run on any but an MS signed Xen binary which MS then bundles with XP. The don't charge for the Xen software since that's GPL'd but they can and do charge for the enabling signature.

Suddenlly MS have a windows platform that runs linux. If the distro has a MS digital signature that is. They get to charge for that too. Probably they'll require linux apps to each bear a separate signature so they can sell or whithold individual apps according to marketing strategy.

Of course ayone else is welcome to sign their own linux distros for Xen - but not for the MS Xen release because redmond won't discose the numbers and the DMCA makes reverse egineering illegal.

Meanwhile - who needs linux? XP now does all that strtaight out of the box! Hoorah!

Sanity check, anyone?

Re:hmmmm...

[glitch!]

The Xen layer wil have some sort of deep down DRM.

I am also thinking that DRM may be involved with this, but from another angle. Microsoft has been hinting for a while about a "trusted" computing environment, and this may be an easy way for them to get started.

With multiple virtual machines running, one can be the standard XP/2000 desktop running regular Win32 apps, and another can be the DRM sandbox running their new generation of "trusted" software. By creating a separate virtual machine, they don't have to break compatibility with current software, while they are quietly installing the groundwork for their upcoming WinXP "pound me in the ass" edition.

How many licenses per machine?

[Glonoinha]

Silly question about activation .. actually not -about- activation, but inspired by activitation :

If you only have a single computer with a single CPU, how many copies of WinXP do you need? That one is rhetorical of course, and the answer is One.

Can you run whatever software on that legitimately licensed WinXP machine that you like, assuming it was also legitimately licensed? That one is also rhetorical and the answer is Yes.

Now install VMware on that machine, WinXP as the host OS. By adding VMware you have not increased the number of CPUs or physical machines. If you created three virtual machines (if you had enough RAM and hard drive space, not a stretch at all) and wanted to run WinXP in each of those virtual machines simultaneously - do you need 1 license of WinXP or four licenses of WinXP (one for the host OS, and one for each VM)?

Granted the activation and active license management in XP may not allow this to happen even if in theory it should be allowed according to the 1 license / physical machine license in the EULA - but swap it with Windows 2000 or whatever ... what are the facts?

I am just curious.

Re:How many licenses per machine?

[steve_l]

I think it is something they havent had to deal with yet.

I am running Win2K+visual studio in one vmware vm right now; I can bring up Win2K3 to run office 2003 when I need to go on exchange. But with the block MSDN license *and* a volume 'no activation' key for XP, office, etc. I get to skip activation.

But imagine if I did have to activate stuff everytime I rebuilt a new VM? Within a month I'd have the activation police complaining I'd activated onto 5+ systems, and that therefore I was violating some license.

Yet at the same time, a VM image, once activated, can be shared and of VMware makes all the hardware look the same, the system doesnt have to think you need reactivation.

So legitimate users of activated apps under a VM will suffer -we have to go through reactivation grief- yet there is now a new way to bypass activation -ship an image of the OS+apps already activated; you just run it in a window.

I dont know how long it will take the 'activation' police to deal with it.

Re:How many licenses per machine?

[questionlp]

Although this (Word document) applies to volume license customers that have their OS covered with an Upgrade and/or Software Assurance, Microsoft is allowing those customers to run two instances of Windows on a single system to allow dual booting and running another instance of Windows under a virtual machine.

I doubt if it applies to the original OEM license since the OEM determines the "product use rights" of the operating system that they include with the computer. I also doubt if that also applies to retail products, but who knows.

Re:How many licenses per machine?

[Anonymous Coward]

The answer is simple. You need 1 real license and 3 "virtual" :)

Re:How many licenses per machine?

[The Wicked Priest]

The best approach is to get the Corporate version of XP that doesn't require activation. ;-)

Re:How many licenses per machine?

[Glonoinha]

Right, or just use Win 2000.

I was more posing the question not how to circumvent it technically, but from a honest licensing perspective how many licenses we are -supposed- to purchase. It is a hazy line, and I really am interested in knowing what is honestly expected by the licensing agent (be it MS or whoever.)

I'm not about to drop $40,000 on server licenses just so I can run 5 copies of Win2003 Server EE in windows on my box, but in a completely compliant world does MS expect me to?

Xen on top of VMware?

[d_force]

Interesting concept.

I wonder if anyone's tried running Xen *within* a VMware image. While this may seem redundant, it could be a method of trying to get the "best of both worlds"... that's assuming the additional overhead is well justified.

-- dforce

Why the hell did sourceforge.net have to go into maintenance mode *today* of all days...?

Re:Xen on top of VMware?

[Anonymous Coward]

Yes, Xen runs just find on top of VMware.

Re:Xen on top of VMware?

[Anonymous Coward]

Yep, it runs, I tried the demo CD on it earlier today.

Just trying to figure out how to get it installed to a hard drive - the CD runs OK, but it refuses to boot on the hard drive after GRUB is installed.

It'd be nice if there was an installation guide of some sort.

It's possible with Virtual PC on my Mac

[GPS Pilot]

Virtual PC lets me copy and paste between Windows and Mac environments.

Re:It's possible with Virtual PC on my Mac

[nacturation]

Virtual PC lets me copy and paste between Windows and Mac environments.

That's because you're emulating another operating system within an operating system and you can bring up, for example, the Windows environment on your Mac. I could be wrong, but I suspect that since you're not running an OS within another OS (rather, alongside it) this wouldn't be supported.

OTOH, I guess there's a really easy way to find out! ISO just finished downloading... :)

Microkernel?

[sleepingsquirrel]

Sounds to me like Xen is a microkernel like thingy (what with the hardware abstraction layer and all) which you have to port your operating system personality to. What would make this different than the port of linux to the L4 microkernel (besides the Windows XP part)?

CAMBRIDGE, HOME OF THE PENIS

[Anonymous Coward]

I agree with this post.

I love you

[Anonymous Coward]

Please put your trousers back on.

NomadBIOS

[Anonymous Coward]

NomadBIOS did this a year ago, including swift migration of Guest OSes betweeen hosts, but with only two people and without a MS-research grant... But Xen looks nice, and the porting effort seems to be smaller.

Re:NomadBIOS

[Anonymous Coward]

Doesn't sound as well tested.
It's been run "on a small network of two machines with hard-coded IP addresses".

Does it actually run real applications with good performance?

How long before SCO makes claims?

[Dcnjoe60]

Any guesses as to how long it will be before SCO makes claims that this, too, is a derivative work of UNIX?

Just more systems to patch

[sfmarco]

Seems just a bigger hassle to me to have multiple Windows Xp running. Every system has to have the security patches applied. Getting my productivity to a near zero.

But on the other hand that might generate more jobs.

Marco

Adeos

[joe_plastic]

http://www.opersys.com/adeos/index.html
http://sa vannah.nongnu.org/projects/adeos

"The purpose of Adeos is to provide a flexible environment for sharing hardware resources among multiple operating systems, or among multiple instances of a single OS."

This is similiar to xen and the new plex86 and it has some experience with supporting rtos

Re:Adeos

[Anonymous Coward]

It's a nanokernel. The "overlying" OSes are not protected or isolated from one another.

Neither... it's a silent X

[spookymonster]

NT

Old skool callout.

[xenoweeno]

The only thing Xen means to me is this.

Interesting...

[Anonymous Coward]

... but I'll wait for M$-DOS port.

Re:I bet it's not Open Source...

[Anonymous Coward]

Why does the original post say:

...and regularly use it for running demanding applications like Apache, PostgreSQL and Mozilla.

but their page now says:

...and regularly use it for running demanding applications like MySQL, Apache and PostgreSQL

????

You bastard!

[Anonymous Coward]

I'll never sleep with you again.

Great Idea! Why all the whining?

[HopeOS]

The x86 architecture is not designed to be virtualized, at least not for code expecting to run in ring 0. The work that VMWare must do to emulate ring 0 without the guest OS coming unglued is simply daunting, including traps for every detail from setting and clearing interrupts to modifying page tables. A lot of that work requires predicting code execution paths and dynamically modifying the code at runtime. It's a real nightmare.

Plex86 promised to provide an open-source solution to this problem, but the last time I checked (which admittedly was awhile ago), they were planning on using bochs to emulate ring 0. This is not exactly satisfactory, but I don't blame them. It's a hard problem, one for which VMWare is entitled and deserving of reward for having solved.

So how do you get real virtual host performance out of your x86 machine? You design around the flaw. Operating systems that are written to run in ring 1 and call the Xen hypervisor instead of performing ring 0 functions will run at nearly full speed. Those that do not or will not make these compromises will see at best, an emulated ring 0. It's really that simple.

Personally, I don't see much value in being able to run any arbitrary operating system in a single environment if it's impossible to get real, sustained performance out of it. Cross-platform testing: sure. Kernel debugging: certainly. But host virtualization? What I want is the capability of running multiple hosts simultaneously, and if the operating system needs delibrate tweaking to make this possible, then obviously that's the direction to go.

To answer the gripes of people who want just that transparent hosting of unmodified OSes, it would be interesting to see a follow-on project for dynamically modifying the guest OS to thunk-out all the privileged calls to use the Xen extensions directly rather than trapping exceptions as they do now. This would probably not work well for page tables modifications, so a hybrid system might be employed. In this fashion, the best of both worlds may be achieved.


-Hope

Yawhn! Yet another RT-Microkernel in disguise

[Anonymous Coward]

From reading the white paper I am fairly diappointed. Xen's entire discovery seems to be that x86 has four rings, two of which are seldom used.

So they force an OS to be rewritten as an application in ring 1 instead of 0 and let their mundane Microkernel run in ring 0. So whats new here? VxWorks, pSos et. al have done this for decades. And rewriting an OS like XP that needs frequent patches is no where near as useful as the authors claim, because every patch needs to be tracked and backported. Given the ample Performance of modern hardware, I think VMware ESX makes more sense by many orders of magnitude .

Why Is This So Hard?

[Chasuk]

Is it really that fucking difficult to write an accurate summary?

A person reads the summary, and reasonably thinks:
"Great! An Open Source equivalent of VMware Workstation! Now I'll be able to run multiple OS's on my Desktop machine without the expense!"

Then he reads:

"Xen requires guest operating systems to be ported to run over it..."

This changes the picture dramatically, and should have appeared MUCH earlier in the summary.

Come on, I know this is only Slashdot, but stupidity and dishonesty like this get really annoying.

GNU/Linux isn't faster than Darwin, just Aqua

[caveat]

I'm using GNU/Linux most of the time on my iBook because:
1) it's faster

KDE/Gnome/&c. is definitely more responsive than Aqua (I wouldn't say "faster", since the UI doesn't really do tasks, it just needs to be snappy and low-latency), and while the monolithic vs. microkernel should make LinuxPPC sans GUI a little faster than Darwin, I've heard here and there (can't back up, sorry) that Darwin has the edge on commandline linuxPPC, probably due to the tight hardware integration Apple can exercise.
Interesting thing i realized while reading this article, all the "robust" (commercially marketed) Unices except BSD/OS seem to run on hardware-software packages (Solaris, IRIX, HP-UX, AIX)...does this mean Apple is going to join the Old Boy's Club?

this looks familiar

[todorb]

from the ADEOS site (http://www.nongnu.org/adeos/):

"The Adeos Project

The purpose of Adeos is to provide a flexible environment for sharing hardware resources among multiple operating systems, or among multiple instances of a single OS.

To this end, Adeos enables multiple prioritized domains to exist simultaneously on the same hardware. For instance, we have successfully inserted the Adeos nanokernel beneath the Linux kernel, opening a full range of new possibilities, notably in the fields of SMP clustering, patchless kernel debugging and real-time systems for GNU/Linux."

this thing is used in the RTAI realtime linux project.

Two Words:

[Anonymous Coward]

fly cheat

(Not that I remember what it is. And now I'm over two words, but I don't care)

Looking at MS Code in Academia...

[anonymous cupboard]

I know you are being funnz but it raises an interesting point. If you have access to the XP source code through an academic license, does this mean that you are 'tainted' from contributing to the Linux kernel?

For example, in earlier days, it was relatively easy to get hold of commercial OS sources without selling your soul (or signing an NDA). For example, I used the source code to write a couple of neat hacks which were available withing the user group.

Are these guys now going to have problems after they have seen at least the lowest levels of XP (I guess, the Hardware Abstraction Layer).

Re:Looking at MS Code in Academia...

[Torne]

No, we aren't. I'm one of the students who has been working on the XenoXP port, and the source licence agreement I have signed does not forbid me from doing related work. I can't copy bits of the Windows code into other projects (or disclose said code) but there's no problem with me contributing my own work to Linux at a later date.

Re:Looking at MS Code in Academia...

[anonymous cupboard]

Thanks, that seems reasonable.

It just sems a pity that you can't produce a series of binary patches to XP so that the rest of us could use this. It certainly would be a great toy to have for developers.

In any case, well done to you and your colleagues. My download (wget not the Bittorrent) is going slowly, so I haven't had a chance to play yet.

Re:Looking at MS Code in Academia...

[Torne]

True. Ultimately the goal is to have a server edition of Windows running on Xen; XP is by way of being a feasibility study. At that point, I'm sure MS will look at distribution terms.

Right now, XP does not get far enough for anyone without the source code to even test in a way that makes any sense, as its progress is measured by many tens of thousands of lines of debugging output, much of it meaningless without reference to the source, so talking about distribution now is somewhat pointless. It certainly can't bring up an actual display =)

Torne
(XenoXP porter)

CPU support for virtualization

[akuma(x86)]

This technology will become significantly easier to implement once Intel introduces hardware support for virtualization (called VT - Vanderpool technology).

IDF Demo

During the Fall 2003 IDF demo, Paul Otellini ran a demo with a Tivo-like app running the Simpsons on one virtual machine, concurrently with another virtual machine booting up windows-xp. One computer was running 2 operating systems and driving 2 monitors concurrently.

Re:CPU support for virtualization

[valleyboy]

Reminds me a bit of Virtual Partitions (VPARS) on HP-UX. Always found it cool to have an N4000 with two virtual partitions and then reboot the one partition whilst the other happily sits there running Oracle!

The main difference with this (apart from the architechture) is that each virtual machine has to have it's own hardware in VPARS, so for example if you were on an 8 way N4000 one possibility would be to have two 4 way virtual machines. You also need seperate network cards SCSI adapters, etc for each VPAR. Very very cool though.

We used this to Split each of the hardware partitions (NPARS) on a Superdome into two VPARS. The Superdome was a DR machine so we split it up to allow it to bhe used for QA and DR. A bit mind boggling to administer but very nice....

Sources are on CD

[anonymous cupboard]

It is true that their 'CVS' tree isn't - it s a BK tree with all the flames seen on the kernel lists. However the ISO contains their complete 1.0 sources.

As for the BK license - you can't use it to develop a competing project but if you separate your work via BK from the work you do on a.n.other version control system, then nobody is going to complain too much.

Intel's fault anyway....

[anonymous cupboard]

If the existing behavior of INtel kernel mode instructions when called from user mode was always consistent (i.e., a trap), then it would be *much* easier to emulate the instructions inside a user-mode VM.

All it takes is for Intel to provide a proper 'V' bit - but frankly, after all these years of architectural holes I do wonder.

.Not umodified distributions...

[anonymous cupboard]

Yes, you can take any distribution but they need to include some assembler macros to replace instructions (I guess, PUSHF and POPF for one) and rebuild the system That is, it isn't quite the original distro anymore and the distro can't self build. Actually, this seems very much like plex86-lite's approach.

great for testing out potential system killing app

[Anonymous Coward]

... or hand coding tweaks in the kernel for that matter! WOOHOO!

Awesome.

[pb]

I'm posting this from Xenolinux right now, and I must say, I'm quite impressed. The virtualization is seamless, the time sharing is quite well done. I haven't tried stressing it too much, just running nbench on one domain (that's how they divide it up) and mozilla on the other (forwarded through ssh back to the first domain). ...now if only I could run some other OSes on here as well...

Doesn't this fsck over DRM?

[adrianbaugh]

I know I'm posting late so no-one will ever read this, but doesn't Microsoft fsck over all their DRM credentials WRT WinXP if they allow a virtualisable version of it to run like this? Surely it is possible to run a virtualised WinXP and a virtualised linux; use the WinXP to decode encrypted or one-time DVDs etc. and use a kernel module in the virtualised linux to read the memory (or swap file) and blat them to TVout, a different format, lineout etc.

VMware tweaking

[Anonymous Coward]

I've always been under the impression that VMware offers the user a chance to optimise the emulator towards a particular guest OS. Does VMware not fundamentally offer one a complete virtual x86 computer with a BIOS, VGA graphics, and IDE drives? Pehaps I'm being pedantic, but you're giving us the impression that VMware could do nothing without choosing a particular guest OS for the host to play to. I agree that $300 is a bit much for a home user, but, uh... a friend of mine gave me a serial number... and I'm thoroughly impressed with the performance and features of 4.x.

Nothing is new under the Sun

[leob]

Just to bring things in perspective:

In USSR, in late 80's - early 90's there was a mainframe of original Soviet design called Elbrus-B, a successor to BESM-6 - the workhorse of the Soviet space, scientific and military computing. To provide smooth transition from the proprietary OS of BESM-6 to UNIX (the sources of Version 7 were available), a so called Dispatcher of Virtual Systems (DVS) was designed and implemented. The feature set was exactly the same. Unfortunately,
unlike BESM-6 that was on par with the U.S. computers of the time, Elbrus-B was way behind, and did not last long, especially after 1991.

DVS had an extremely tight and elegant interface, and porting both operating systems to it was quite straightforward. Xen is pretty much the same thing for x86. Therefore I dub Xen the Dispatcher of Virtual Systems for x86!

Xenify, anyone?

[leob]

With all the achievements in x86 binary code analysis, instrumentation and compilation (think of valgrind or Transmeta, to name just a few), wouldn't it be possible to read the kernel binary into some kind of a compiler, find all uses of the instructions that have to be replaced, replace them (and the adjoining code, if necessary) with the Xen interface calls, then reassemble and relink the kernel?

How extensive and how non-trivial are the necessary changes?

Xen was the worst

[t0ny]

I dont know about anyone else, but Half Life really got annoying once you get to Xen. And what was up with that stupid big-head boss thing?

errr..are you joking?

[Anonymous Coward]

That way, when Ingo Molnar decides to start using 80286-style 16-bit tasks with lots of grow-down and conforming code segments to do system calls in Linux 2.6, we won't get caught out too badly.

Joke? Or is there a legitimate use for this 16-bit crud in Linux?

Re:errr..are you joking?

[Anonymous Coward]

Joke.

kma is making fun of OS/2's use of crazy ia-32 features, if I'm reading him correctly.

He's a hard dude to read sometimes, though.

Re:I bet it's not Open Source...

[Torne]

The current XenoXP port is built on top of XP Embedded, which does not require product activation; activation is done at the time you assemble the disk image just by providing an Embedded licence key (no phoning home). Actual licensing terms for XenoXP will be decided by MS (under the terms of the Shared Source agreement, we may not distribute XenoXP to anyone other than other Shared Source licencees) as and when they decide to; this will not be soon as right now, the port is not complete enough =)

Also, you don't really have virtual hardware in Xen, not the same way VMWare does; we don't emulate hardware. A virtual network interface is nothing more than a ring buffer, likewise for disks.

Torne
XenoXP porter

Re:I bet it's not Open Source...

[Elwood P Dowd]

(waiting for the 20 second delay to expire... @#%$! Slashdot)

In IE, right click on the "You're posting too fast, you asshole" page, choose "refresh."

It'll ask you if you want to repost the same form info. Say yes.

That way you don't have to retype your post, and you don't have to wait another 20 seconds. Other browsers I've used don't blank your forms when you click back, so this technique isn't as necessary. So I don't know if it works in Moz or Safari or whatnot.