Windows beats RedHat in web server benchmarks

[n0cmonkey]

Here's the report.

 

[episodic]

Yea, says Microsoft.

 

[Zugzwang152]

Originally posted by: episodic
Yea, says Microsoft.

actually, says a non-microsoft benchmark...

 

[n0cmonkey]

Originally posted by: Zugzwang152

Originally posted by: episodic
Yea, says Microsoft.

actually, says a non-microsoft benchmark...

Sponsored by...

 

[Zugzwang152]

Originally posted by: n0cmonkey

Originally posted by: Zugzwang152

Originally posted by: episodic
Yea, says Microsoft.

actually, says a non-microsoft benchmark...

Sponsored by...

on such an easily repeatble test, why does everyone doubt the authenticity of the results?

 

[CTho9305]

Some slashdot comments said they used different encryption for the two servers, with the Windows server using a type that is 3-4x faster. I don't know if that's true.

 

[kamper]

Why are we debating a year old test? The redhat and apache versions they used are long since gone (well, not the apache version, but there is much newer).

 

[Zugzwang152]

Originally posted by: kamper
Why are we debating a year old test? The redhat and apache versions they used are long since gone (well, not the apache version, but there is much newer).

wow i didn't even see the 2004, i read May 5th and assumed it was this year. :Q

 

[akugami]

Well...considering how many "independant" tests we've seen favoring MS over the years I'd say we have every reason to be doubtful. It's not so much the tests they're running that favors their OS but the fact they try to pass it off as independant.

 

[n0cmonkey]

They used Apache 1.3.23 with RH AD, and Apache 2.0.40 with RH 8.
Despite the fact it's a load, Apache 2 is supposed to be quite a bit faster. They only use RH 8 on the dual processor machine (I think). Maybe TUX doesn't work with apache 2.

They're also using kernel 2.4.9-e.3enterprise when testing 2, 4, and 8 processor machines. 2.4.9-e.3 with the uniprocessor machine.

Is 64k bytes the default block size for NTFS on a volume of 60GB (if that matters)?

p33 mentions the encryption stuff.

Operating system: Key exchange algorithm, encryption, message digest
Win2k3: RSA, RC4 (128-bit), MD5
RH AS 2.1: diffie-helman, 3des (168-bit), sha1
RH 8: diffie-helman, 3des (168-bit), sha1

rc4 is not only faster than 3des, it's absolute shit. I'm not sure how RSA compares to DH, but they're both good systems.

 

[n0cmonkey]

In some really unscientific openssl testing (openssl md5 some_random_pr0n.avi), md5 seems to be a bit faster than sha1 also.