As far as comparative costs go, the determination really comes down to the local context in which a company selects its server. There are clearly situations in which Linux would be a preferred cost savings for one company while Windows 2003 server would be more cost-efficient for another. That was the gist of what I saw revealed in the abstract of the article I posted.
As far as a comparison of their security features goes, that really comes down to the competency of server administration in my opinion. A good server administrator running a well-thought-out and competently executed security setup on Linux will be superior to a less competent and less well-thought-out system on Windows 2003 server and vice-versa.
The real issue around which a comparison of Linux and Windows 2003 server should be based is that of the cross-platform capabilities of the two and, with .NET installed on Windows 2003 server, it is no contest, Windows 2003 server wins hands down. This is what I pointed out in
my first post on this thread and is something that has been verified from the real-world testimony ImaGraftedBranch supplied in his
post #176, in which he pointed out that he was working in a situation in which he was pulling data from multiple platforms of a wide variety supplying over 70 types of datasources. Only .NET can do this with ease.
There is nothing that runs on Linux, or anything else for that matter, that can compare with this level of cross-platform interoperability. That is why the charge so many make against Microsoft that it is not "economically democratic" -- which was the original issue to which I responded in this thread -- fails the test. The great myth of open source software is that it supports an IT world in which everyone can communicate with everyone else. It doesn't. The companies supporting Linux are working very hard to exclude data access and data exchange with Microsoft systems. That is the reality of open source software.