How good is Vista?

I’ve had the opportunity to do some in-depth poking around various Vista installations. Not just those being deployed onto the desktops of company and corporate users, but also the installation you get when purchasing a new home PC or laptop, as well as the sort of installation you can end up with by upgrading your home PC from Windows XP. It’s a very mixed tale, which doesn’t help much when the internet is full of claims and counter-claims over just how “good” Vista is, and how reliable the installation and upgrading process is in practice – some people are swearing by it, others are swearing at it. To be honest, it seems Vista is getting off to a poor start with a vocal minority, so as an attempt to get to the bottom of this matter and take a temperature reading of what’s going on, I’ll report my findings in a variety of scenarios.

Let’s start with the business desktop, because this is the easiest. First, we need to remember that business desktops typically don’t come with an OS installed: the company will have a rolling installation licence from Microsoft and will normally generate its own install images. The comparison here between XP and Vista couldn’t be clearer. Back in the bad old days of XP, you had to make a separate install image for just about every flavour and sub-flavour of hardware you had in your company, then patch it with the necessary drivers and applications, too, plus keep it up-to-date in a process that consumed man-weeks of work. The end result was that XP images tended to get done once a quarter if you were lucky with your time resources, but more likely twice a year. Or being more realistic still, a new image was created only when new hardware came into the company, and it wasn’t touched again until a major service pack came along. The end result was that the OS image was merely a workable, if somewhat feeble, starting point, and installing it triggered an avalanche of further downloads, hopefully from a local Microsoft update server but more often over the internet. Want to complicate the issue further? Have a network that spans various European countries, because typically each XP image worked for one language only. Managing the inventory of images across such a WAN was frankly a horrible job, and my heart goes out to those who’ve had to deal with such messes.

With Vista, it’s almost all completely different: for a start, a single image can hold the information needed for all the hardware in your organisation; it can be multilingual and contain all the applications, too; better still, it’s trivially easy to patch the image files to keep them up-to-date. Ideally, I’d prefer a Windows server applet that just sucked the updates out of the Windows update server and blew them into the images on a daily basis, but maybe that’s something to hope for in the future. Nevertheless, even with what we have today in Vista, this whole deployment issue has become a breath of fresh air. Or, to be a little more cynical, Microsoft has finally come up with the solution we should have had all along, and it’s entirely its own fault that we had to put up with that XP mess.

For many companies, this improved deployment will be more than enough to justify the move to Vista. The savings in time and manpower required to keep the images up-to-date is significant and shouldn’t be overlooked. Unfortunately, many companies work on the basis of “cough, mumble, mumble” when it comes to OS imaging, and they get hard stares from me when I find them out. Bad imaging is the root of considerable problems and distrust in their Disaster Recovery (DR) plans. Often their DR plans are made on the basis of “how long would it take to get this shiny new Dell up and running?” while ignoring the fact that it’s the first one in the building and that every other machine they own has its own version of the image disk, now wildly out of date. If you have to respond in a real DR scenario, it’s highly unlikely this particular Dell (or Compaq or HP or…) will be what you’ll buy. You’ll end up with whatever you can get your hands on quickly, at which point your imaging strategy has fallen to pieces and your DR plans fall with it. Keeping up-to-date OS images should be seen as part of the DR process, not as part of the deployment cost.