Hyper-V release

After about 3 years of coding and development (a lot of which was, by all accounts, carried out by XenSource before the acquisition by Citrix), Microsoft have finally released Hyper-V, their long-awaited hypervisor and management system. I haven't seen it myself, and wouldn't actually know what to do with it if I had, so I'll just concentrate on the commercial impacts I believe this will have.

According to reports I have read and people I have spoken to though, the first version of this product isn't exactly up to much with regard to features and functionality. But then it doesn't really need to be. It will be baked into Windows Server 2008 and will be essentially free so it's unlikely to ever really set the world alight. Having said that, I'm sure it'll be good enough to do the basics of virtualisation and thereby you will be able to enjoy the benefits you get from this process - consolidation of servers, more dynamic data-centres, reduced need for power, cooling, space etc. - by using Microsoft products instead of a competitor's. This was actually, in my opinion, always one of the main drivers behind Hyper-V. Doing VMWare some serious damage. More on that in a second.

So, alongside hurting the dastardly VMWare (who have successfully managed to really rub the guys at Redmond up the wrong way), why have Microsoft gone down this route? How are they going to make any extra money from it? Internal discussions with Dominic and David here at COMPUTERLINKS led us to 3 main reasons for doing it:

1) The hypervisor itself will cost $28. Times that by the amount of server licenses Microsoft sell and that's not a bad start, particularly as the Hyper-V edition of Windows Server 2008 will be the default. But I'm sure Microsoft are expecting improved revenues of the management suites around it too. Things like SMS, MOM, Virtual Machine System Centre and other management tools as well as the advancements around VECD (Vista Enterprise Centralised Desktop or virtual desktops), which ideally need a virtualised datacentre to function properly, will also add incremental business in no small fashion.

2) The world is moving that way anyway. Let's not forget, Microsoft are way behind the curve on this. They failed to move early enough on the wave of virtualisation, allowed VMWare to build up a substantial lead and are desperately trying to catch up now. The business reasons for virtualisation and dynamic datacentres are robust enough for this to take off in a big way and Microsoft simply can't get up amongst the leaders early enough.

3) I saw a documentary on Bill Gates a couple of weeks ago where Sir Alan Sugar was also interviewed. He referred to the Microsoft sales guy who came to try and sell him an operating system all those years ago as a "transatlantic smoothie". He also refused to buy MS-DOS for his Amstrad computers until the cost came down so far that it was as good as free. Microsoft (and Gates admitted this freely), were doing nothing but seeding the market. Any income from it was good but the main point was that, if a customer was using Microsoft products, that meant they weren't using a competitor's. Google had a similar approach. Offer your utilities for free for as long as it takes for people to become dependent on them, then bring in the charges - mainly in their case to organisations who want to advertise to those people using your utilities, but also (I'm sure the future will bring this) to those people themselves. Ironically enough, Google, and the whole concept of cloud computing, probably now offer the largest threat to Microsoft's current domination of global IT, but that is a whole new topic for another day.

So where does this leave Citrix then? They have spent the last 13 or 14 years practically selling Terminal Services CALs for Microsoft, spent 500 million dollars last year on a hypervisor and are now in the unenviable position of having to get a return on that investment just as their biggest partner brings out their own competitive solution. (All this "co-opetion" speak no longer floats my boat I'm afraid. It's a competitive product, just as VMWare ESX is.) And the big difference this time around is that Citrix don't have the protection of the TS CAL story to help them out.

Imagine the scenario. You've bought Windows Server 2008 with Hyper-V. It doesn't quite do everything VMWare does but you've saved your company many thousands of pounds at the same time and can live without the schnick-schnack. Why would you then decide to buy Citrix to "add value" to your Hyper-V? With Terminal Services it was different - that's mission critical so the proposition that Citrix offered on top of TS was justifiable. Virtualising servers and managing them isn't really mission critical. It may become so in the future but it's not right now. Brian Madden thinks it will take 2 years for VDI to become mainstream and also, interestingly, thinks Citrix will actually ditch XenServer altogether.

I wouldn't perhaps go that far but I also have my doubts. Can this nice, cushy, you-scratch-my-back-I'll-scratch-yours relationship with Microsoft last much longer?

Postscript 01/07/08: Apologies, I misunderstood Brian Madden in the article he wrote to which I linked above. He is not predicting the end of XenServer, he is predicting the end of Citrix' support for Xen, the open-source hypervisor, as he now clarifies in this article.