Is the time rapidly approaching when OpenVZ becomes accepted as the second player behind VMware -- or at least, the one that isn't Microsoft? Or is it destined to fall into the gap between two increasingly polarised camps?

In the virtualisation technology market right now, there are two main bidders for second fiddle behind clear market leader VMware -- and they're both open source projects: XenSource, which develops Xen, and SWsoft with OpenVZ.

From the point of view of many, Xen has been the favoured platform until quite recently. It carried an open source halo, was heavily linked with Linux -- Novell has already shipped it with SuSE and Red Hat's community version of its distro -- Fedora Core 5 -- includes it too. It's widely expected to be part of Red Hat Enterprise Linux 5 -- the next release, unless Red Hat and Xen fall out completely.

What's more, Xen promises much. It works with a para-virtualised Linux kernel -- one that's been altered to pass certain OS calls through to the hypervisor layer. This technique offers high performance virtualisation that's almost indistinguishable from running an OS on the bare metal.

But then XenSource recently teamed up with Microsoft, with the aim of adding Windows support to the previously Linux-only Xen. This makes perfect sense from XenSource's point of view: Microsoft is the only competitor to VMware that looks at all likely to be able to meet VMware on its own terms. Even though Microsoft's enterprise-level hypervisor, which will compete with ESX Server, is not likely to appear before 2010.

And you can't ignore Microsoft. Although it started in the virtualisation business with its usual approach of selling commercial software, there are clear signs that it's prepared to change its ways in the virtualisation market, where the open approach -- in various guises -- prevails.

Anathema to open source devotees it might be, but Microsoft boss Steve Ballmer said recently that Microsoft is no longer opposed in principle to open source. After all, there's too much money at stake.

Whether or not Microsoft does so, the Xen tie-up means that Microsoft proprietary code is bound to make its way into the GPL-bound Xen product. Such mixing of licensing approaches could cause difficulties when it comes to distribution -- although Microsoft has powerful lawyers on its side who'll probably be able to bolt together a compromise.

Where does this leave newcomer SWsoft with its open source project OpenVZ?

Project manager Kir Kolyshkin reckons that his product is ahead of Xen for a number of reasons. "We use a single kernel and so we are lighter, while Xen has multiple different kernels," he said.

"With VT enabled hardware, you can run Windows and Linux on same box, also BSD," he said. "We have a single kernel to run virtualisation stuff to create those containers. Our overhead is lower than Xen too -- it's about two or three percent."

Kolyshkin says that its strength also lies in its alliances, such as deals with IBM and -- critically -- with Red Hat, whose executives were quoted last month as saying that they didn't think Xen was robust enough to go mainstream -- even though this was semi-retracted a few days later by an announcement smothered in marketing-speak.

Many have since pondered whether this might represent an opening for OpenVZ to supplant Xen; its technology, though not identical to that of Xen, is similar.

However, OpenVZ could be in danger of falling between two stools, as the market polarises around the two main camps: VMware plus numerous partners versus Microsoft and Xen. And it's Xen that OpenVZ can expect the most problems with, as Illuminata analyst Gordon Haff says: "With so many free, near-free, and built-in product out there, it's going to be increasingly tough for a standalone virtualisation company to make money -- especially if you look out 18 months or so."

While it's going to be hard to make money out of virtualisation products that aren't on the cutting edge, the strength of conviction that many open source advocates bring to their work should not be underestimated. As Kolyshkin says, when asked why SWsoft set up the project: "It has a value of its own -- we want to contribute to Linux and have it become a better OS."

Virtualisation is too young yet both as a technology and as a market to descend into an oligarchy. With so much development and market potential still ahead of virtualisation technology, it seems probable that OpenVZ has a future. Whether it's one as an independent company or as part of a larger portfolio -- EMC-style -- remains to be seen.