Theres such a blizzard of variables that accurately calculating the costs may come down to your individual situation.
When it comes to choosing between Windows and Linux for an enterprise, there are seemingly endless studies one can turn to about the total cost of ownership for either operating system. Yet one thing needs to be remembered, no matter which way youre migrating: the human cost.
Windows and Linux both require training and experience to use welland this goes doubly so if youre migrating from one OS to the other and need to jettison existing work habits and acquire new ones. The TCO studies that Ive seen typically mention these things as part of the system support and administration costs, such as admin salaries, and so are not always obvious. To that end, I took the time to think about what these costs really mean, and how they often manifest in the real world.
The first people cost, and the most obvious, is having people on staff who can properly administer whatever platform youre choosing. The exact cost for this sort of thing varies wildly between the different TCO reports Ive studied, partly because everyone has a different way of quantifying that cost.
Some will claim that Linux requires less people cost because once its set up it tends to remain running without a great deal of nudging on the part of the admin, and therefore requires less on-call time and other administrative overhead. Some would put the cost at about even, if only because the ease of deployment for Windows tends to be balanced out by the ease of maintenance for Linux as cited above.
Finally, some cite Linux as being slightly more expensive in this regard, not just for the cost of the in-house expertise needed to run it but for the possible cost of external support and consultancyneeded either to get things running or keep things running, depending on the scenario.
With both Windows and Linux administration, though, what matters is having people who can respond creatively and intelligently to problemswho, if theyre uncertain about something, will take the time to educate themselves about it and come up with long-term solutions. Such people are never cheap and shouldnt be treated cheaply, because they can stay with you for the long haul.
Another factor is the cost of retraining existing users to work on the new OS, which is often not so much overlooked as misinterpreted. One area I see this extending most into is what applications are being used. If youre moving from Microsoft Office to OpenOffice and most of your workers are not doing anything particularly adventurous with the program, the total amount of retraining needed probably wont be too big. Retraining users for workaday processes usually doesnt cost very much. Outside of that, though, if there are specific, technical things that youre training users fori.e., writing macros or performing other work with the program beyond simple daily usethen youll need to factor in the cost of training and research for those things.
The cost of retraining can also vary widely within a single organization, simply because not everyone will need to be retrained in the same ways. To restate the above example: if youre in an environment where most of the people use a heavily macro-driven set of Word documents, the bulk of your userswho just use Word and dont code for itwould not need to be retrained if you switched to OpenOffice.org. However, the few folks you had on staff to create those macros would need to be retrained on the new suite.
The exact amount of time needed to retrain is usually not very long for basic functions. One study I looked at about retraining Windows users on Linux concluded that 80% of the people who switched to Linux needed only about one week to retrain themselves to use Linux properly for their needed job. That does raise the question about how long the remaining 20% took to adjust, and even if it wasnt a significantly greater length of time, it does indicate that retraining isnt going to be a one-size-fits-all job. (I should also point out that its possible to switch to OpenOffice.org entirely apart from Linux, and that in fact many people have done just thatremained on Windows but used OpenOffice.org to lower TCO, at least initially.)
In the end, though, the only truly accurate and useful TCO study for comparing Windows to Linux is the one you conduct yourself. You can look to other TCO studies to get an idea of how to model one for your own organization and needs, but youll need to know your own environment intimately to get a real grasp of the cost. Whenever possible, look to another company whose work habits and needs match your own and learn about what they didyou may be surprised at how the costs added up or leveled out for them.
Sometimes you need to spend a lot of money now to save money in the future, and what looks like a bargain at first doesnt turn into a bargain later.