Some days ago, RedHat claims that Xen is not ready yet for production environment. Now, Novell strikes back and defend their inclusion of Xen in their latest SUSE release.
That's so funny, because in fact they don't give any concret argument. RedHat says that it's not mature, and Novell just answers that they've contributed to the code.
As a member of the Debian Xen team, I'll arbitrate between us.
Xen is nowadays deployed in a lot of solution and in a lot of system architecture. I don't think this is because the Xen technology is mature, but only because it's the solution which fit the best the need of virtualization (and cost rationalisation) that organizations may have.
Xen can not pretend to be stable, mature, or whatever you want, because it is still not integrated into the Linux kernel tree, and thus far for a good reason: the API/ABI is not yet stable.
I know that in theory it is, but in fact it's quite wrong and you cannot rely on that specification. Xen works fine but from a developer point of view it is not in a stable and fixed shape.
It seems that Linus and others kernel people are not very happy to integrate directly Xen into the mainline kernel, and I can understand them, since this technology is rather new. I don't think Xen will be a suitable solution which can be deployed for years until Linus decide to integrate a virtualization system into the Linux kernel.
Furthermore, Xen is still a huge source tarball of mess with a big kernel patch.
And even if I really think that Xen will be the adopted solution for everyone, it won't be in its current state.
So, I think RedHat is pretty right. Just think about the Xgl/AIGLX stuff, that's quite again the same behavior between this two companies.


