Slashdot Mirror
VMware: Another Netscape?
An anonymous reader writes "
This CRN article states that Microsoft is about to buy Connectix and enter the server consolidation market. Connectix makes virtual machines products that compete with those of VMware. Quote: 'The technology will be integrated into the Windows code, sources said.' Will Microsoft be able to pull this one off? Will their virtual machines run operating systems other than Microsoft's?"
...now we can have a tail-recursive win32 delay loop.
Trusted Computing FAQ | Free Dawit Isaak!
if virtual pc will be suspended for the mac.
are they more concerned about stopping adoption of os x, or more concerned about selling windows licenses to mac users?
If you are running 50 instances of NT Server on a single box, how many NT licenses do you need?
It's most likely desireable because they want to be able to run partitioned servers, much like one can do now with VMWare. Of course, I'm sure they won't mind 'embracing and extending' the product out of Mac-Space. It is probably the core virtualization technology that they are after though.
that they're just trying to find some way to make it look like typing "ls" on a Linux shell gives you a BSOD.
According to MacCentral. This could be good for the Mac, meaning the development team would have more access to Windows code and be able to guess how things are working less. Or it could be bad. And I have no idea what to think. Microsoft still makes money off of the license that goes with the sale of VirtualPC.
Of course MS will buy one of the implementors of this kind of technology. Look at Citrix. Of course, it will run well, 1 or 2 versions later. Of course, it will NOT run other OSes as well, or even at all. There will be undocumented hacks, which might make it work better.
The problem is that MS stuff doesn't run on anything but x86 these days. I want a real hardware platform, like IBM makes, where I can carve out a few LPARs on a 32-way box with 8GB of RAM. Then I'll run Windows200x on it, with my other OS in that. Real hardware redundancy, etc.
Using Linux as an example--
Its far better to run Linux and Win-in-VMWare (free + VMWare) than MS and Linux-in-its-VMWare-clone. Do you trust MS stuff to be the core OS?
I want to delete my account but Slashdot doesn't allow it.
Comparing this to the browser battle isn't a good example. I doubt that MS will allow other OSs to run, thus VMWare will still have the market for running Non MS OSs on Win2k/XP. Plus, I doubt that MS will offer any functionality where you can run a MS OS on top of a non MS OS (although they may, since they'll still sell licenses in that situation), thus VMWare keeps that market too.
VMWare isn't going away. They just may take a hit on the running multiple Windows on Windows market.
'The technology will be integrated into the Windows code, sources said.' Will Microsoft be able to pull this one off? Will their virtual machines run operating systems other than Microsoft's?
This is most likely Microsoft's response to Solaris Containers which are expected to be shipping in Solaris 10. Of course, both of these are simply implementations of ideas pioneered by IBM with VM/CMS.
The VM approach makes a lot of sense even if you only plan to use it to run multiple copies of the native OS within them. The advantages are twofold. Firstly, it prevents one malfunctioning application from impacting other applications - even on Unix this is a serious problem, since one process can devour the CPU, memory, disk space, etc. Secondly, it allows resources to be redistributed or added on the fly, especially if your VM is seamless enough to span nodes.
What's wrong with this? Now maybe we can finally get a PlayStation emulator built into Windows.
I wonder if this is part of an attack against Apple?
As those of you not familiar with the Mac Marketplace might not know, Connectix makes the popular Mac application Virtual-PC. Virtual PC allows Apple owners to emulate a complete PC enviornment on their Apple machines, at somewhat reasonable speed.
They seem to have had favorable licensing with Microsoft in the past, as they offer pre-installed images for certain OS systems, such as Windows XP, 2000, etc. While they do (I assume) pay MS for each license, it does help people to break the MS dependance gradually, as they can still run their old applications under emulation.
If they eleminated this crutch for people switching to apple, and then later discontinued Office... Apple would lose most of it's corporate market.
So- As useful as this technology is in the Server market (and keep in mind this is closer to Bochs than VMware), I can see MS execs encouraging this buyout to help keep control over the future of Apple.
Colin
Colin Davis
I think MS's biggest problem is they try to clump too many things into one, that and a companies hierarchy can scale only so much. Trying to add something like this that is extraneous for the most part is just going to screw things up. It's not a criticism of MS, I don't think anyone could do what they are trying to do well. It's simply too much.
"Not knowing when the dawn will come, I open every door." - Emily Dickinson
they SELL windows liscenses??? Gee, and I thought the .txt file with the key in it was included with EVERY copy of windows...
who knew???
of coure, I better not let the BSA hear me say that, they might give me a merit badge in thievery!
This is my sig. Its pathetic.
After all, each copy of VPC represents another Windows license (for the most part). I would think it would be to their advantage to get as many mac users as possible using VPC.
Or maybe it's just a way to extend the Windows monopoly, and maybe DRM/Paladium/etc. A few years ago, I was in a store where a customer was returning an iMac, complaining that it was constantly crashing. Turned out that the user ran VPC full time, and didn't know what the MacOS was.
I use Macs to up my productivity, so up yours Microsoft!
It has been stated that VPC for Macintosh is now under the control of the Macintosh Business Unit. [MBU]
The biggest problem with this, of course, is the fact that Bill and Co. may just decide that the only application that the MBU needs to push out is VPC. This means no more Office X, no more native X applications, just run the Windows version of the app in VPC [slowly and painfully]
Oh man this is bad news. I wonder if the DOJ even cares.
Blocklevel: Practical Information Architecture
Johnny has 25 NT boxes and 19 of them have WVMS (windows virtual machine software) running on them. If 7 of those 19 are running WVMS within WVMS and 3 of those 7 are running Win2k Advanced Server, and the other 4 are running WinXP Pro, while the rest of the 19 are using WinME in the WVMS for backwards compatability issues. How many licenses will you need?
Bonus: How much will this cost including the inflation of the economy and of Microsoft's prices by the year 2004?
It is kind of curious that Microsoft would choose Connectix's product for its virtualization. For those not familiar with it -- Connectix Virtual PC is a little more elaborate than VMware because it actually emulates the i386 CPU in software. This is why it works, for example, on a Macintosh, while VMware doesn't.
Now, it should be patently obvious that Microsoft doesn't want you running Linux-on-Windows, Windows-on-Linux, Windows-on-Mac, or anything other than Windows-on-Windows. So you have to wonder what they're up to, here.
When you don't have cross-OS stuff to worry about, why emulate the hardware? For that matter, why emulate a computer at all? For Linux-on-Linux applications, you probably won't choose VMware when you can instead run User Mode Linux -- it uses the hardware more efficiently, you can share filesystems between the host and virtuals using NFS, and it runs the host OS's native binaries. I would think Microsoft would prefer to go this route.
Or perhaps Microsoft has finally decided that Itanium is an ongoing disaster and they need an Intel exit strategy? Hmmm...
Tired of FB/Google censorship? Visit UNCENSORED!
I saw the article already, but based upon M$ history and the announced integration of yet another application into already bloated and non-secure mess that Windows is, I foresee future news, with a familiar flavor. I.e. "this exploit allows anyone to take over any instance of blahblahblah".
Yeah, they also said they would continue to support Mac computers, but is this something you really want? I couldn't help, but notice a comment that 'they don't intend to kill the software'. Really... It's just one more sword to dangle over Apple, when Steve gets too uppity.
I don't see any long-term winners here, other than those selling Connectix's assets.
A feeling of having made the same mistake before: Deja Foobar
This stuff is really cool, and I'm glad the industry is starting to wake up to the possibilities.
I see this as something that's more likely to popularize virtual computer technology, rather than something that's likely to eliminate our options. Obviously, I don't have a crystal ball, and I could be wrong.
I have a box that I use mostly to run VMware client OSs. Linux is my host OS, I have a very sparse and clean linux from scratch system set up on the box. I've got all kinds of stuff stashed away in various VMs.
The great thing about this sort of setup is the flexibility. The client OSs are basically just data files on the host os. If you copy the files, you've backed up the system, or cloned it.
You can move the files to other machines that have different hardware -- you don't have to worry about the sound and video card drivers.
And you can even replace the host OS without being too disruptive. I used to run redhat as the host OS, but I copied off the data files, set up my linux from scratch system, and brought the data files back in. Everything was fine.
The result of this is that the chains of dependency that exist between hardware, operating system installations, and applications become much less restrictive.
Another result is that it's trivial to play with new systems -- I don't run OpenBSD, for example, but everytime they could out with a new one, I install it, just to keep my hand in.
All this is, at bottom, is just a more flexible way of looking at OSs. An OS becomes a blob of data that's easier to move around from one hunk of hardware to another. And it's easier to keep lots of those OS blobs on a given machine.
It's a great way to deal with "staging" servers. You can take a production server (which is really a VM), copy it, and do whatever you want to the copy, without damaging anything. When everything is working properly, you can slide the new server into place. If you need to revert, you can just go back to the old data.
I suspect that this functionality is part of what MS is after.
He's finally working on a way to port the BSOD over to Linux and OSX for us!
yes, if you read of any Connectix's docs you'll see that they emulate an S3 Trio 64 video card, as well as the rest of the hardware. I have installed RH 7.2, 7.3, 8.0, Mandrake 8.0, 8.2, 9.0, Lycoris (Redmund), Suse 7.2, 7.3, 8.1 with success.
First let me say that I really like VMWare. I think they did a fabulous job - it's one of the better-engineered pieces of software that I've ever seen. All the times that I've had the pleasure to use it, it's worked for me without a hitch, despite the subtle complexity required to do an application like that well.
Furthermore, I hope that one day we'll see a real, meaningful government reform at Microsoft that puts them out of the business of "innovating" away various application markets.
My needs for VMs have been sparse. Most often I'm testing something (like an installer) that sprays stuff all over Windows, and it's just simplest to roll it back using the Undoable disk when the test is over. Or maybe I've got some code I want to check out that I consider really dangerous. Once in a while, if I'm stuck running Windows, but I need a Unix service on the network for a little while, I can raise a virtual linux server and keep it running as long as I need it. Far more convenient than hauling out another box.
I can see the attraction in virtual machines. You have so much more control. Bluescreens don't hang everything - only the particular virtual CPU they happen on. And VMWare's code is so freaking efficient, I can play counterstrike with a few of these virtual servers running, answering queries in the background. But it seems silly for virtual machines to become institutionalized in that role. To me, that's evidence of failure in the OS design. You have a reliability problem? Fix it in the OS. You have a control problem - something you wanted a VM and Undoable disks to solve? Add a feature to the filesystem. You have a security problem? Definitely an OS issue.
VMWare et al are great for ad hoc stuff and I think sooner or later most developers would be glad to have it around, but if you plan on running it all the time, in a server environment for instance, then it's just a big kluge. Your OS wasn't _designed_ to run inside itself... it's a big resource waste. Fix the problems in the OS. Compartmentalize, if that's what the environment demands. But don't do it this way. It's just goofy.
Want to Know How to Cheat the GPL? Read On!
It's so that customers can run NT4-specific apps under future products like Windows Server 2003. See this eWeek article.
Four months ago, our company tried buying a copy of VMware with WinXP licenced to run in the VM. VMware said that they were working out a new license with Microsoft so they could sell XP and that we should call them back in a couple of months. Our purchasing guy has called them once a month since then and we still can't get it.
Now I know why it's taking so long...
I've been playing with User-Mode Linux a bit recently; it's a port of Linux to run on Linux :-) (instead of running on real hardware, it does hardware-ish things via Linux syscalls). It runs as an unpriviledged user, but has its own internal users, permissions, even a root user.
It's a nifty idea, but it's not suitable for servier virtualization in the data centre, at least not yet. The problem is that the host Linux kernel lacks resource allocation and accounting capabilities - other than say nice there's no way to really manage the CPU, and you can't quota the network bandwidth in and out of the VMs, you can't limit the working set size of each VM, and so on. A process misbehaving in one UML VM can still affect others on the machine.
The real use for UML is in development environments, it allows you to very quickly set up test systems. Start 5 VMs and now you can test your distributed app for race conditions without having to buy and spend time configuring physical kit.
Microsoft would have no reason to want to stop VPC users buying windows, at all, no.
However, Microsoft also has no reason to want certain things about VPC to stay the way they are. For example, the fact it is screamingly fast. For a long time, one of the big bragging points mac users had was that we could run windows, *emulated*, at about the speed as a windows machine with half the mhz. (I don't know how current models perform.) That's really, really impressive insofar as emulation goes. Microsoft also has no reason to want VPC to continue to be as clean and effective as it has been.
What i am saying is that people don't come to VPC on a lark: it is an expensive piece of software, and people come to it becuase they need to get something out of it, usually to run some windows-only program. This means VPC's quality can suffer, and Microsoft will have no reason to consider this a bad thing-- at the moment, VPC has no serious competitors, so people will keep buying VPC.
Microsoft also has no reason *not* to stop Virtual PC from being able so cleanly, seamlessly, and easily to emulate, say, Linux. They have no reason to make it easy to run a non-MS operating system on your mac.
There is also no reason not for Microsoft to continue as they have and then, after a couple versions, slowly let wierd bugs, incompatibilities, etc, creep into VPC., until mac users *still* can run windows, but they only do so becuase they need to run windows for some reason-- because VPC has become enough of a pain that the PPC's wonderful talent for emulation no longer seems like much of an advantage over the x86.
Am i saying Microsoft is going to do this? Well.. no. In fact, i don't think they will, becuase macslash is reporting that apparently the VPC team will report directly to the MacBU, not to seattle. This means that they will continue, almost certainly, to make VPC as much a quality product as possible. So there goes that conspiracy theory out the window right there.
However, it does bother me that Microsoft is able to take big, important groups like Connectix and Softway (Interix) and buy them up just like that. Yes, they are buying them for apparently benign purposes. But what it seems like to me is that while Microsoft is not buying these companies so they can quash or disable them, they are buying them so that they can keep their eye on them. Potentially, something like Interix or VPC could become a big stepstone in some kind of major migration away from Microsoft. if Microsoft owns those companies, however, if it looks like such a thing is going to happen, MS can take steps to prevent it, so long as MS always keeps the quality of those companies' products so high that there never is a reason for a competitor to arise. Threat management.
This brings me to my question: how on earth is MS going to make Palladium work with VPC? Palladium becomes pointless unless those keys are kept secret, and if MS embeds those keys into a macintosh executable then extracting them will be trivial. So how is MS planning to make Palladium work in VPC? Are they going to require a PCI card with a palladium chip in it, or what? That would still toss out Palladium's concept of the secure keyboard-to-processor-to-monitor path, but it would at least keep the keys locked safely in silicon. Or, much more likely, are they just going to not let VPC run palladium apps, since the Mac OS is not "secure"?
So, here's a slightly more likely conspiracy theory. Perhaps MS [only partially of course-- i've no doubt they're mainly buying Connectix for the reasons they say they are] likes the idea of buying Connectix because it removes the risk Connectix will attempt to emulate Palladium within VPC? I mean, Palladium is going to be damned hard to crack, but if anyone at this exact moment in time has both the resources and the reason to crack palladium, it's Connectix or nobody. I really haven't the foggiest idea what Connectix was planning to do about Palladium, but they have experience at cracking closed systems-- they reverse-engineered the PSX. That expertise, and a few hours rented time with an electron microscope to pull on the Palladium's keys, and suddenly MS is no longer the sole source or vendor of their Palladium platform.
Would that have actually happened? I have no idea. But it certainly won't now. Maybe not a big deal, but certainly convenient for Microsoft either way, no?
Just like it's "convenient" that Bungie's excellent cross-platform game development library, rather than being sold off with Oni and Myth, is currently buried somewhere deep in the bowels of the earth..
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
For all the Mac paranoid out there, here's another conspiracy theory that will undoubtedly turn up on the rumor sites.
;-)
Microsoft really wants to revive, sell, and support an improved version of Connectix RamDoubler for the latest revision of Mac OS 9. Heck, they might even make it into Microsoft RamTripler (MS Ram*er for short)
Connectix stopped selling RamDoubler and promised to do away with support for it in September 2003. But an killer utility like MS Ram*er will cement the last Mac holdouts to Mac OS 9 and their old hardware.
The conspiracy has widened!
Do you remember the last time when Connectix sold it's Playstation emulator to Sony.
There was lots of talk about how good this would be, since Sony could cerate an official platform for selling PSX games to mac and pc users, since Sony was supposed to lose money on consoles this would make perfect sense. Of course, this did not happen, Sony chose to kill it instead.
Now, why do I get the feeling that the exact same thing will happen again?
Somewhere in the heavens... they are waiting.
Almost everyone here seems to be missing the point. This is not for home use. This is not intended for you, Joe Schmoe Windows at Home user to run other operating systems.
This is for the server market. We have an IBM mainframe at work that is currently running approximately 6 virtual machines. Not so that you can play a Windows game in Linux, but so that the mainframe can offer more services. Although I do believe that one of the virtual machines is a fairly standard installation of Linux of some sort, every other OS on the system is a very specifically tailored OS for a specific job. I'm not the administrator for this box, so I can't say too much. But I know that there are specific Tivoli UNIX versions installed, as well as an TSM/ADSM (backup) specific OS.
I think that THIS is what the article is getting at. This is not about you playing Tux Racer on your Windows box.
Sig.i>