Weekly Edition
Return to the Press page
| |||||||||||
|
| |||||||||||
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch)
Linux-Watch looks at the
release of paravirtualized network and block device drivers that allow
Windows Server to run unmodified in Xen virtual environments. "These
device drivers support SUSE Linux Enterprise Server 10 (SLES), and work on
Intel-based server platforms featuring chipsets using Intel-VT
(Virtualization Technology). The new drivers will let customers migrate to
newer and fewer energy-efficient servers, consolidating legacy Windows or
Linux solutions onto virtual servers."
(Log in to post comments)
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 13, 2007 21:50 UTC (Tue) by mmarq (guest, #2332) [Link] Virtualization ( help)
Since software virtualization is generally implemented via a layer of virtualization software, sometimes referred to as a VM monitor (VMM), that presents the illusion of many VMs...
How came is Xen going to be integrated with KVM/paravirt_os ?
Isnt Xen a VMM hypervisor ?
Isnt KVM a VMM layer ?
I can see the advantage of having each VM as a Linux process, instead of delegating in a cokernel faction, but will it be possible to KVM to maintain the same domain "live migration" ability of Xen ?,... that is, can KVM guest domains be also a Xen domain ?
And the split driver model of Xen, that is, the paravirtualized driver model... would it be necessary ?,... that is, since the VMM is *in* the Linux kernel itself why cant a paravirtualized OS, a Linux guest as example, use directly Linux drivers just like another process ?
If a qemu implementation is going to be integrated too, i remenber reading something of that kind somewhere, does it mean that a VM can comprise a mixture of paravirtualized devices and emulated devices, at the discrection of the user ?
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 14, 2007 1:23 UTC (Wed) by k8to (subscriber, #15413) [Link] This is going to sound rude, mmarq, and I apologise for that, but I'm going to say it anyway.
I feel you are making a habit of posting your thoughts-of-the-moment which relate only tangentially to the article. I don't see anything about KVM and Xen being integrated in the article. Is this speculation on your part?
Speculation on articles can of course be lots of fun to read, but I generally come away from your posts with no idea what you are trying to communicate. Perhaps it is because you often raise 3-4 issues in a single post. Maybe if you clearly delineated the start of each new issue?
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 14, 2007 11:13 UTC (Wed) by evgeny (subscriber, #774) [Link] Second that...
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 14, 2007 4:46 UTC (Wed) by dlang (subscriber, #313) [Link] the various hypervisors are incompatible with each other and with clients designed for each other.
there have been proposals to change this, but they were not accepted (good old NIH from every other virtualizaion project, among other things)
however this really isn't relavent to the topic of drivers being available for windows to let it run on a Xen modified Linux system.
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 14, 2007 15:43 UTC (Wed) by mmarq (guest, #2332) [Link] "" however this really isn't relavent to the topic of drivers being available for windows to let it run on a Xen modified Linux system. ""
I agree and apologise.
But it could be relevant for something like ReactOS on KVM/Xen/Linux. Run Linux drivers instead of windows ones.
Intel, Novell deliver virtual Windows drivers to Linux (Linux-Watch) Posted Feb 14, 2007 15:48 UTC (Wed) by k8to (subscriber, #15413) [Link] ReactOS could already run anywhere it wants to. It's open source, and so can be readily modified to run where desired.
Trying to dig through the marketing talk..... Posted Feb 14, 2007 22:04 UTC (Wed) by jmorris42 (subscriber, #2203) [Link] Ok, first I was wondering why a paravirtualized Windows Server needs the hardware virtualization enabled processors. Then I noticed how they were careful to only say they are working on paravirtualized network and block device drivers. Looks like they have decided to just pick the low hanging fruit and let the hardware virtualize the video, mouse, etc and paravirt the performance critical disc and net.
Looks like a big step forward at any rate, especially if they release these drivers as open source so it could be ported to support other paravirtualizion systems, etc.
And of course it had better work with both Intel and AMD's hardware virtualization. But I won't frag em too hard for 'forgetting' to mention their arch enemy in their press release.
|
|||||||||||