Slashdot Mirror
VMware Opens Up API to Partners
mstansberry writes "This week VMware opens up its source code to its x86 partners, calling it the best mix of open-source and proprietary. While the general public won't get a look at the source code, the likes of IBM, HP, Red Hat and others will. Releasing an API is a way for a company to bring more people into the fold and to get more applications integrated within the platform. But from the looks of last quarter's financial reports, VMware doesn't need much help getting people on board."
...will jump on the vmware bandwagon. With Mac OS X and Mac OS X Server (especially Mac OS X Server in the context of what I'm about to discuss) supporting x86, it would be trivially easy to have Mac OS X Server run in a supported fashion in a vmware environment on any variety of hardware. Stay with me for a moment: similar to the impact of soon being able to get an Apple desktop or, especially, laptop system that runs Mac OS X plus any x86 OS, including Windows, in a sure-to-exist virtual machine/vmware-like environment at near-native speed of the underlying hardware, having Mac OS X Server run on vmware in a server environment - somewhat the reverse - would be a huge coup for Apple in the datacenter.
Yes, yes, we all know that Apple, at least at the outset, will not "allow" Mac OS X to run on non-Apple hardware. Aside from some semi-insane but actually interesting prognostications from John Dvorak (and TPM panic aside), Apple is primarily talking about the desktop/consumer marketplace when it says this. There is little to nothing to stop Apple from partnering with an enterprise x86 vendor (or a partner such as vmware) to provide a vehicle via which to run Mac OS X Server on hardware other than Apple's 1U, single-power-supply Xserve.
Mac OS X will only run exclusively on Apple hardware as long as its good for Apple. As soon as it becomes desirable to allow Mac OS X (or Mac OS X Server) to run on possible non-Apple hardware configurations, you had better believe they'll do it. That's probably part-and-parcel to this whole x86 transition strategy. Further, consider that individual market segments may be appropriate for this first, such as enterprise datacenter and server markets. Consider also that while Mac OS X is $129 ($69 government and education), Mac OS X Server is $499/$999 ($249/$499 government and education), meaning that Mac OS X Server has a price point much more in line with allowing Mac OS X Server to run sans Apple hardware and still be a profit center. And as it matures, Mac OS X Server is an increasingly powerful, very attractive UNIX server platform, with major commercial vendor support and the best of the open source world wrapped up into one product.
I see Mac OS X Server on (something like) vmware on non-Apple x86 enterprise server hardware in Apple's future.
Good, I hope one of these companies has an interest in modifying VMware's BIOS so it can boot an OS from a USB device (USB-FDD/USB-HDD/USB-ZIP). That is one missing feature that really irks me.
How does VMWare fit into EMC's strategic interests? I'm surprised that Apple didn't buy VMWare when they had the chance because with the move to x86 hardware, having VMWare part of OS/X would be killer.
So by "best mix of open-source and proprietary", they mean "not open-source at all".
In case you were wondering, VMware is an application that lets you run several virtual machines on one host machine, and even set up a virtual network of those machines and bridge it to the real world if you want, allowing honeynets and the such.
I hate headlines that list some alphabet soup without explaining what the heck it is. I read about 2 years of RSS headlines before seeing an article that mentioned what RSS was.
Bruce
Bruce Perens.
I think that VMWare is finding itself in potential trouble because it is not going to be easy to sustain their financial success with the Open Source projects such as QEMU and Xen gaining ground.
I personally think that hypervisors are overhyped (pun fun!), and that the most practical and useful form of "virtualization" is actually separation as is achieved by Solaris Zones, FreeBSD jails and (the most advanced of them all IMO) Linux Vservers. A pretty good article on it here.
Separtion carries nearly zero overhead, simplifies administration because there is one kernel and one filesystem. It allows for simple "entry" into a virtual server from the main server, and there are other subtle advantages that I can't think of right now probably....
I believe it will come bundled with Duke Nukem Forever
Correct me if I'm wrong, but couldn't the better machines from the old mainframe days virtualize themselves? IIRC someone ported Linux to one of the virtual machines and was able to run a crazy number of instances on one set of hardware.
I know this sounds (is?) crazy, but why not open up the architecture of the old mainframes, and base the next generation of PC hardware on those ideas? CPU and memory are cheap now, but those suckers were designed to be robust back when you couldn't solve all problems by making large clusters of faulty machines - they had to work, period. Surely modern PCs could match much of the power of an old mainframe machine, properly designed, and the whole modern desktop OS and apps could just be run on a virtual instance from a PC. This would allow, say, Windows and Linux to coexist, run at the same time, have no issues crop up that software like VMware has to work around, and allow for all sorts of interesting debugging possibilities (how about booting up another VM to debug a wiped out desktop OS, just by pressing a button on the keyboard?)
IIRC x86 has some real issues with virtualization, but if what I have heard is true and x86 is now mostly a layer put on top of more advanced cores in most CPUs perhaps the problem has already been (largely) addressed. Does this makes sense to anyone else - would it be good to have "desktop mainframes"?
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
"But from the looks of last quarter's financial reports, VMware doesn't need much help getting people on board."
It's got nothing to do with revenue, it's about market share. Virtual machines are going to be huge in the coming years, especially in the webhosting market. Pretty soon, leasing a "dedicated server" will be simply leasing a dedicated "instance" of a server, for lack of a better word.
EMC wants to keep their lion's share of the market, especially with products like MS' Virtual Server 2005 and SWSoft's Virtuozzo entering the fray.
I did some contracting work for Big Blue a few months ago, and their deployment teams LOVE VMWare. They used it for all kinds of crazy stuff, and it worked amazingly well. I hadn't used VMWare since a very early beta back in the 90's, and was impressed at how well it has come along since then.
EMC is just protecting its market share now as best it can, before others start chipping into it.
If you don't want to pay for VMWare, I would suggest trying out QEMU.
Ever since the "QEMU accelerator" module has been released (version 0.70), it has worked as a virtualizer as well as emulator, so you can get almost VMWare-like performance (that is, if you just want to run Windows under Linux or vice versa). QEMU itself is licensed under LGPL, the accelerator module is free as in beer (and there's another, open-source accelerator project in the works, though I'm not sure what the situation is today)
AMD will release Pacifica Q1'06 which provides OS virtualization in hardware.
This will allow Xen to transparently virtualize linux, windows, macosx, etc.
What you are seeing is VMWARE desparately trying to entrench themselves in the virtualization market before Xen & Pacifica (and whatever Intel's processor is) makes their product technologically non interesting.
From what I understand they offer nice add tools, and that's pretty much the only way they can actually have any sort of future in the virtualization market.
With the open source term becoming the latest wall street Good Thing buzzword, the risk-reward ratio of releasing source code, more often than not in a very limited way, is an attempt on behalf of these companies to declare to public investors that the integrity of their product is strong enough to the degree that they have no compunction of releasing their secret blueprints, essentially inviting people to come hack pbrush.exe and VMware. To hackers, most of this OS releasing going on is of software that has no hacking appeal. There's just no motive to capitalizing on these OS releases.
I, for one, am not impressed by what strike me as PR maneuvers which at best are patinas devoid of true significance worthy of a meaningful press release. Otoh, I suppose it may increase shareholder wealth, the legal purpose of a public company.
VMWare ESX is NOT Linux...
We just tried (and failed) to virtualize four machines on a single dual-Xeon system with 3GB of RAM using VPC. After dicking around with it for a week, we switched to VMWare and had it running well in about an hour. Save for the shoddy documentation, VMWare trounced it in every way.
I don't think I'd consider VPC a competitor to VMWare today. Next year, maybe. Right now? Not from what I can tell.
Dewey, what part of this looks like authorities should be involved?
For many VMware is just workstation product. Because that is what they get to see people discussing in the LUG. Please checkout www.vmtn.net (VMware technology network) and see the discussions on ESX,VC and ACE from the enterprise users. The feature set that you get there is mind blowing. I cant think of anything equivalent in any other product/OS currently or in the near future. VMware has many "first" to their credit that no other software provides/provided. But I would say they have been such a low profile company. They are hitting the headlines only after their EMC acquisition, which is understandable considering they are moving more into Enterprise segment. Just a partial list of features: NICteaming across different NIC make and models at the kernel level Virtual VLAN Beaconing NIC Failover PXE boot SAN Multipathing Multi vendor SAN support at the kernel level SAN path failover Hot backups of virtual machines through redo logs VMotion (move VMs from one physical host to another without the underlying OS knowing about it) Perl/COM APIs to control Virtual Machines Multiple level of snapshots with VC Cloning from the same base image in WKS By far the largest number of guest OS support. ACE -Virtual machine deployment (http://www.vmware.com/products/desktop/ace_featur es.html)
The product and its features just speaks for themselves. Go and read the specs or try it for yourself.
Do not compare VMware with Xen/Virtual PC or any other projects. They have just started to do things that VMware did 6 years ago.
yes pacificia and vanderpool will let anyone do virtualization. So what ? If vmware could do so many things when there was no hardware support for virtualization, imagine what they could bring in when the support is built into the hardware.
With them already ahead of the game by miles, I could only see that vanderpool and pacifici help them proliferate further in the server market space.
Finally, I'd like to point out that Xen is close to zero overhead for most system level benchmarks. Due to licensing restrictions (which I think are not entirely unreasonable) on VMware prodcuts, I don't have numbers for VMware's overheads. Intuitively, though, fooling an OS into thinking it's *not* in a VM requires more effort than not fooling it - VMware will always have to do more work than a paravirtualised solution like Xen, so it necessarily incurs more overhead (for now).
Whilst I'm about it, I should also mention some more things that are under development. Yes, you can always say things are "on the way" (and I'm sure VMware have cool things in the pipe too). Nonetheless this should be arriving in the foreseeable future and since it's an OSS project it's not a secret...