Credits: R.E.Ballard r.e.ballard@usa.net

I wonder if there are any other benchmarks in support of Linux ???!!!

There are other benchmarks, but they are written in German. In Germany, using NDAs to restrict publication of unfavorable benchmark results is illegal. It is considered a form of Fraud.

If not.....why NOT!..ask yourself this question before deploying Linux.

If you are a member of MSDN (which most really knowledgable NT administrators and developers are) you are not allowed to publish evaluations, benchmarks, or even bug reports without Microsoft's written permission. Microsoft doesn't give this permission to negative reviews, benchmarks, or bug reports.

There have been numerous incidents where CERT alerts were reported without Microsoft's permission and the security problem was not published until Microsoft released a patch. In the case of attacks such as Back Orifice and Explorzip, Microsoft released ineffective patches and forbid the disclosure that these patches were ineffective.

http://www.microsoft.com/ntserver/nts/exec/compares/ntlinux.asp

I know what Microsoft are like... that benchmark would be 100% fabrication!!!

We really need to clean up our language :-). Profanity doesn't impress most people, and it alienates the managers and decision makers whose influence is most desired.

It would be silly to expect Microsoft to provide pointers to negative reviews. Look at their coverage of the DOJ, and compare it to coverage provided by ZDNet, ABC News, Netscape, and the raw DOJ coverage (mostly pure depositions and transcripts).

For "equal time", you have access to other publishers. Microsoft uses the EULA, and MSDN agreements to restrict publication of certain information, especially "Reverse Engineering".

They publish things like that because they are scared (explitive deleted) of Linux... and who needs to benchmark Linux?

Benchmarks are very interesting. The Mindcraft benchmarks were deliberately rigged to give NT every advantage possible.

Put simply, it was similar to running drystones on a dedicated NT running only the drystone benchmark, then running the same benchmark on a Linux machine that is running hundreds of "default processes" in the background, and has 20 users attached. The NT Dhrystone would be faster because Linux is constantly interrupting the benchmark to do important work.

There are many numbers Microsoft will not allow it's customers to publish:

If Ebay crashes, they can't point at NT. If Merril Lynch pays $10 million in SEC penalties due to an NT crash, they can't mention Microsoft's role. If Sure-Trade is inaccessible for hours at a time, the only clue that it is an NT system is the use of .asp pages.

we already know that it is better than anything else.

The problem is that corporate executives do not have a great deal of experience with Linux. Many don't even realize that half their file-servers and most of their e-mail and internal web sites are running on Linux. In some cases, they think that they are still running NT and assume that it is NT that is so reliable.

Unfortunately, it requires a court order to get the real information. It's trivial to get the syslogs of UNIX and Linux systems and the event log of NT systems to see how often systems were rebooted. It's even pretty easy to import those into spreadsheets. The story is interesting. NT systems usually have to be rebooted several times/week. Linux systems get rebooted about once every 3 months.

Currently, there is very little accurate accounting of Linux activity. Based only on number of "official" units shipped and sold, Linux had 17% of the market and was growing at 217%. NT had about 30%. What wasn't covered was the number of times that a Linux system could be legally replicated, the number of "free" (uncounted) CD-ROMs included in Books.