Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?

Comment: Re: Ok.... (Score 1) 48

by ls671 (#48674689) Attached to: Kodak-Branded Smartphones On the Way

It apparently is indeed. I always suspected that this was due to the simplicity of English and my belief that you can write it "almost" properly with little knowledge of grammar and analysis. In contrast, to write in French "almost" properly, it's hard to get away without knowing those rules. When used to grammar and analysis, the "it's vs its" dilemma becomes easy to solve.

Comment: Re:Why bother? (Score 1) 418

by ls671 (#48651645) Attached to: Ask Slashdot: Is an Open Source<nobr> <wbr></nobr>.NET Up To the Job?

I have seem at least 10 benchmarks showing the same. Do your own research and try to be a little less biased. Latest apache does the same as nginx now. Event driven model with multiple connections handled by a single thread. That is as simple as that and it shouldn't be too hard to understand that using the same techniques, apache will indeed be as fast.

See a more detailed benchmark here: (search for nginx for comparison)

Comment: Re:Why bother? (Score 1) 418

by ls671 (#48645539) Attached to: Ask Slashdot: Is an Open Source<nobr> <wbr></nobr>.NET Up To the Job?

> both in ease of administration and more importantly, in performance (why is Apache still spawning processes for every request that comes in..

Apache has had thread model (no need to fork new processes) and EVEN nginx like event model, with fewer/threads than connections for quite a while. Update your info. We run latest IIS and apache on production servers and IIS seems to do pretty well but saying the things you say about apache isn't fair and false in some cases.

In fact, we run instances of apache with event model (max 75 threads, max 4096 connections, keep alive 5 seconds) as a caching reverse proxy and WAF (mod_security) in front of IIS application servers and others. Centralized WAF configuration for all applications and caching so all static content requests rarely hit the application servers.

App server behind the caching reverse-proxy get 2 to 15 work threads, keepalive rarely on, the config is fine tuned with the help of benchmarks and yes, some backend apps are faster with only 2 workers and often degrade with more than 15.

This is a typical performance setup.

If an experiment works, something has gone wrong.