Linux Study: Microsoft Credibility Questioned
by Steven J. Vaughan-Nichols | linux-watch.com | November 2005
A New Linux Study Suggests Fundamental Microsoft Creditability Problems

Another day, another lame attempt by Microsoft to show that Windows is better than Linux.

This time around, Microsoft commissioned a study to show that Windows does a better job of serving e-commerce applications than Linux.

Of course, in the study, they didn't use the same e-commerce or back-engine DBMSs.

OK, right there, without saying another word, anyone who really knows anything about benchmarking knows that the study is fundamentally flawed. You're not comparing apples to apples; you're comparing apples and oranges.

It would be a different story, if you were trying to compare the transaction speed and reliability of e-commerce packages, but that's not the case here. Microsoft was trying to prove that Windows was better than Linux.

To do this "study," Microsoft hired Security Innovations Inc.. Paul Thurrott, a Windows journalist, describes the company as "highly regarded."

I prefer to use Security Innovations's own description of its relationship with Microsoft: "Security Innovation is a certified Microsoft partner for security services. We have both the Microsoft SWI and ACE certifications as an authorized professional services provider for Microsoft technologies."

What kind of idiots does Microsoft think we are, anyway?

In the, cough, study, which compared Windows Server System and Novell Inc.'s SUSE Linux Enterprise Server (SLES), they simulated both the aforementioned e-commerce applications and an upgrade from Windows 2000 to Server 2003, and SLES 8 to SLES 9, and a year's worth of running, from July 1, 2004 to June 30, 2005.

What did Micros... oh I mean Security Innovations, find out?

Well, first, that by Microsoft's own admission the sample size of administrators was too small to provide conclusive statistical comparisons!

Is this amazing, or what? In the executive summary, Microsoft admits that they don't have real data!

So what conclusions did they reach?

They found that with Linux you could solve problems in a variety of ways, instead of one true, Microsoft way. OK, that's true enough. But, this, this is a problem?

Sorry, Microsoft, I don't buy that paying your prices for your integrated innovation solutions is any kind of real business win.

Go call me a capitalist, but I prefer open-source's competitive product approach to Microsoft's "our way or the highway" communism.

The study also found that Windows was dramatically more reliable.

Really?!

That's not the Windows I know. Server 2003 is a lot better than W2K, but in my experience, and with the companies I know, SLES still stays up longer than Server 2003.

You know, I also recall a few potential Windows security show-stoppers over that year. There was the SMB (Server Message Block) over TCP/IP exploit, and a whole slew of holes in TCP/IP -- and those are only a few of the ones that Microsoft has fixed.

Despite that, the study also found that the patch rate on Linux wasn't quite five times higher than Windows. The testers found that SuSE had 187 while Windows only had 39.

Hey, they finally got one right!

Yes, Novell, like any serious Linux vendor, fixes all its problems as fast as possible. Microsoft doesn't. Even when a problem is a potential system killer, sometimes the boys from Redmond drag their feet.

Oh, and funny this, but the SuSE patches tend to work, unlike some Microsoft patches like two recent critical Internet Explorer patches, or the infamous Windows 2000 patch that blew up ASP (Active Server Pages) pages that were running IIS (Internet Information Services).

Microsoft also claimed that Linux patches took twice as long to apply and broke applications.

What nonsense!

In my office lab, I run a W2K server, two Server 2003 servers, and a pair of SLES servers. As it happens I also, during this last year, updated a W2K server to Server 2003 and one of the SLES servers from 8 to 9.

On those systems, I've also installed a variety of server applications including SQL Server and MySQL.

You know what? First, the Linux patches always, always installed faster. And the only breakage I ever saw from either the Windows or the Linux systems was when I was working on W2K.

Do you know why I support Linux over Windows? Because I don't just write about operating systems. I actually use them, and Linux works better than Windows does.

Lest you think I'm only saying that because I know Linux better than Server 2003, think again.

I literally wrote several hundred pages on Server 2003 in an online reference guide to the operating system. You can see the most recent edition of that over at InformIT.

No, I know Linux. I know W2K and Server 2003. And the people who wrote this "independent" study of both certainly didn't know Linux well -- and I have my doubts about the Windows side, too..