Open Source Vs Windows  

I know this subject is controversial, but here's what I have found in 25 years of experience. I use Windows, Linux, Unix, OpenVMS, FreeBSD and other operating systems all day long, and I am experienced with IIS and Apache. My conclusions are based upon actual testing and benchmarking, not rumor and articles by others.

Stick with the brand names for mission critical systems. Open source is fine for many things, but excluding the "big products" such as Mozilla, Apache and so forth, I am leery of it.



I would never even consider open source for the important things such as payroll, general ledger, accounting, warehouse operations and SCADA. These are the applications which companies actually use and need to stay in business.

I have never found anything that even approaches Office XP feature-for-feature. Office XP is rock solid, doesn't crash and is very secure. I cannot say the same for Staroffice, Wordperfect and the other competitors.

Windows 2000 and Windows XP are unmatched in quality as far as their market is concerned. Windows 2000 server is rock solid stable, has an incredible number of features and, most important of all, is supported and well documented.

Windows 2000 also has a security model which is unmatched in the industry. This model came from the VAX (Digital Equipment) and Novell, both of which were (and are) excellent. Linux and Unix don't really have a security model in comparison (I am referring, of course, to active directory and NTFS).

Of course, Windows has the well known issue of security exploits (a different issue than the security model). I currently manage quite an extensive farm of IIS servers, and I've found it's not that much work to keep these systems completely up-to-date. We just have to do our jobs as administrators.

IIS and Apache are equivalent in functionality.

IIS performs better than Apache (I've done the testing myself on the same machine) for straight HTML pages. IIS also has a better security model than Apache by far (based on NTFS as it is). IIS is also a heck of a lot easier to use than Apache, although you can certainly purchase GUI's for apache which make it usable for the average person. And the first major problems (including security) with Apache 2.0 have started appearing, and they are just as nasty as anything found in IIS.

To my way of thinking the main reason to prefer Apache over IIS is (a) initial cost, and (b) knowledge of your people. If you and your group already know apache, then that is the best choice for you. If you already know IIS, then that is probably the best choice.

Browsers? IE won the browser wars for a good reason (besides ruthlessness) - it's far superior to Netscape 6 and before. Now that Mozilla and Opera have had a couple of years, it looks like they might give IE a run for it's money. Even now, though, I've found IE is superior to the competition and I'm sure there is a new version in the works.

But all of this is not as relevant as the cost of modifications. I've found the initial cost of the software, regardless of what it is, is puny compared to the cost of modifications. I would never even consider hiring programmers, for example, to modify the OS, the browser, or the web server, so the availability of sources is not relevant. As far as an application system is concerned, I am interested in modifying the business rules, not the application itself. In fact, if my accounting system requires me to modify it's primary code just to change a business rule, I'll find something else.

There are many companies which provide good application systems, including SAP, Ultipro and hundreds of others. To date, not one open-source version of these (the really important things) has even made it to the initial RFP.

I find the "religious wars" about this subject fascinating. Ask Apache admins and they will tell you Apache is best and death to all unbelievers. Ask IIS admins and you will get the same story. same with Windows, Linux and OpenVMS and any other similar thing. I personally believe we all like what we like and are used to. What I personally like to do is ignore the hype and rumors and so forth and do my own analysis. And yes, I ignore Microsoft hype as well ... they have more PR skills than any hundred other organizations. But PR does not make for a good product.

The decision to use or not use a product should depend upon rational facts, not opinions and noise. Apache is in many ways better than IIS, and IIS is in many ways better than Apache. Which is better? Depends upon what you are using it for, the goals of your organization and many other things. Same with Windows Vs Linux, or any other argument.

Our experience, and keep in mind this is for corporate intranet web servers, is that the cost of Windows and IIS over a long term (5+ years) is far, far less than the cost of Linux and Apache. We performed our own internal study and factored in the cost of the software, maintenance costs, upgrades, time to support, training and so on. Windows 2000 server came out far ahead of Linux (even with all of the patching) and IIS came out ahead of Apache. I have only just installed Apache 2.0 on a windows platform and so have not formed any opinions on this product yet. In a few months, perhaps, I will have some solid data about what appears to be a great new release of Apache.

Cost, however, should never be the primary criteria of this kind of decision. Cost is the least important of many different factors which go into a decision. I have found over my career that when I chose something based upon cost as a primary factor I wound up with that sore behind feeling. Yet when I chose something based upon features and requirements and a good analysis, then picked the product which fit the best, then and only then was it comfortable sitting down.

I have found that Apache (with a good front-end GUI) is superior for a hosting solution. I like Apache in this environment because it gives the users (webmasters for small and medium sized shared hosting web sites) more flexibility without the server admin needing to get involved.

Linux is a great server platform, although we have found the cost to support, train and maintain is higher than windows 2000. As a desktop, though, Linux isn't even on the same planet as Windows XP. Linux has a long way (and I mean light years) to go before it is anywhere near Windows XP as a desktop solution for corporations.