November 2nd, 2005
Windows vs. Linux? It's Apples and oranges
Paul Murphy’s recent article, "The better part of better, faster, cheaper," concludes with this observation:
There’s a moral to this story: that it’s the combination of expertise with technology that counts, not just the technology. The objective reality is that there isn’t anything that Windows Server OSes do "out of the box" that Linux can’t do better but the practical reality is that Windows set up by an expert works better than Linux set up by a beginner -and the scary thing is that most of the expertise out there is Windows only.
I continue to be perplexed that industry pundits continue to compare Windows and Linux as if they were comparing two similar products when in fact they are comparing apples and oranges. One really needs to look at Linux as compared to UNIX to get a true sense of TCO as it compares to capabilities.
In such a comparison, the winner will almost always be dependent upon the scale required. As long as the available horsepower on commodity hardware is sufficient to do the job, Linux will usually win out (due mostly to reduced hardware costs); however, when an application requires the scalability (or stability) only available on "big iron," a proprietary UNIX solution — such as Solaris or AIX — usually has the edge.
Returning to the discussion at hand … Windows vs Linux:
You are certainly correct that the combination of technology and expertise is what counts. You are also correct when you say that "there isn’t anything that Windows Server OSes do ‘out of the box’ that Linux can’t do better" — while UNIX falls at one end of the scalability spectrum, Windows (even the "Enterprise" version of Windows) falls at the other. Despite Microsoft claims to the contrary, Windows just doesn’t seem to scale very well.
Unlike many of your readers, I don’t consider the price of an OS as being any kind of a useful measure of TCO. For instance, most readers will say "But Linux is FREE." Well, yes, if you are a hacker setting up a server in your bedroom, Linux is indeed free (to download) — and is totally unsupported. If you want to buy Linux with media and documentation, the price with limited support starts at about $35. Fully supported Linux workstation software comes in at around $180 — annually. Fully supported Linux server prices start at around $350. These prices are virtually identical to Microsoft’s OEM and retail prices for Windows XP Pro and for entry-level Windows Server 2003. Since the hardware is identical, the "cost" argument is simply moot.
Expertise, expertise, expertise. The entire thrust of your argument is almost lost in your article but it is the linchpin of the entire Windows vs. Linux debate:
How much does someone need to know to install and run a Windows server?
How much does someone need to know to install and run a Linux server?
Virtually anyone with a high school diploma and few hundred dollars can get certified in Windows server technology in a matter of weeks or months — while most people entrusted to administer an enterprise level UNIX/Linux server have a college degree and/or years of UNIX/Linux experience.
How is this possible? By deciding to favor ’simplicity of use’ over ’simplicity of design’ Microsoft has leveraged the low cost of commodity hardware to produce cost-competitive software which is complex in design but easy to use. The result is inefficient (and sometimes bug-laden) software running on over-powered hardware. Throw in affordable training and certification programs and Microsoft has a winning combination.
For the small business, where scalability is not an issue but human resource are, the inefficiencies don’t matter because the overriding factor is the up-front cost of the system and the cost of hiring someone with the minimal experience necessary to take care of the system.
The point that is missed by many outside of the enterprise is that the enterprise measures everything in total cost of ownership and then compares that TCO to productivity. This is a measure almost impossible to make outside of the enterprise itself.
It is in the enterprise that scalability and OS efficiencies come into play. From a purely technical standpoint, the differences in efficiency between UNIX and Linux are very small indeed — and are attributable almost entirely to the relative maturity (and respective development models) of the two operating systems.
So, how does Microsoft compete in the enterprise? Well that depends. For services which are being delivered cross-platform and which require scalability and reliability, not very well. In this environment, Microsoft is being squeezed at the top by UNIX vendors and at the bottom by Linux vendors.
For proprietary services being provided to Windows clients, there is little choice but to go with Microsoft — again because simplicity of use for minimally-trained employees outweighs the inefficiencies of Windows server technology.
In the end, the enterprise will rarely be a purely Windows, purely Linux, or purely UNIX shop. They will choose their platform based upon their immediate need for scalability and stability, the availability of well-trained human resources, and TCO.
Your final comment is that:
… the scary thing is that most of the expertise out there is Windows only.
It’s true enough (for the reasons I have state above) that there are more Windows techies out there than UNIX/Linux techies, but "scary" doesn’t seem appropriate. It might be "scary" if you run a large UNIX/Linux shop because it means that as your staff ages and retires, it will become more and more expensive to replace them with employees with similar skills and levels of expertise. This is simply a fact of life with which any manager who is not a baby-boomer will need to deal — regardless of the industry. Still, with three-to-five year hardware lifecycles, today’s decision need not impact the decisions you make tomorrow.
The fact that the expertise is overwhelmingly Windows today should not be "scary" at all. Remember:
- First it was CP/M, then it was DOS, then it was MacOS, now it is Windows.
- First it was WordStar, then it was MultiMate, then it was WordPerfect, now it is Word.
- First it was VisiCalc, then it was Lotus 1-2-3, now it is Excel.
UNIX predates all of these operating systems and applications and perhaps Linux and its applications will one day displace Windows. UNIX was develop as a highly portable OS which could run efficiently on relatively inexpensive hardware. First DOS and then Windows surpassed that feat. More likely than not, something will come out of the blue to change the IT landscape once again. Either way, sooner or later, someone will displace Microsoft.
| 11/02/05
