TCP stack throughput

Joel Reymont <>
Tue Jul 5 00:17:52 CEST 2005


Matthias,

I ran your C server and I should say that I'm very unhappy with the  
results. Unhappy for Erlang that is. The test runs about twice as  
fast and I don't get any connection resets running 1000 simultaneous  
connections on Mac OSX 10.4.1.

Granted, I get 0 as the factorial result and the C calculation is  
probably way more optimized than the Erlang one but you can still  
connect at least 2 times as many clients as we managed to connect  
with Erlang on Solaris and 3 times the number I managed to connect  
myself.

I invite everyone to run the test and experience the difference.

     Joel

On Jul 4, 2005, at 11:34 PM, Matthias Lang wrote:

> I'm with Mickael.
>
> Your results could be interpreted as showing one _or more_ of the
> following:
>
>    a) That your test client is the bottleneck
>
>    b) That your test server is the bottleneck
>
>    c) That the OS/tcp stack isn't really designed for the sort of
>       use you're testing.

--
http://wagerlabs.com/tech





More information about the erlang-questions mailing list