Title: RE: socket multiplexor

I have just some references:

http://www.kegel.com/c10k.html
http://ninja.cs.berkeley.edu/dist/papers/events.pdf


_________________________________________
Eider Oliveira - UOL
ICQ#: 116119057 * Work Tel#:  +55 11 3038-9492
________________________________________________
Um homem com um rel�gio sabe que horas s�o; um homem com
dois rel�gios nunca est� seguro. (Lei de Segal)


-----Original Message-----
From: Henri Gomez [mailto:[EMAIL PROTECTED]]
Sent: segunda-feira, 13 de janeiro de 2003 12:36
To: [EMAIL PROTECTED]
Subject: Re: socket multiplexor


Eider Silva de Oliveira wrote:
> Hi,
>
>    I don't know if it is the case, I think the best performance solution
> to achieve high scalability (10K connections per cpu unit) will be
> several sockets being handled by an unique thread, in a way we will have
> fewer running thread to attend more sockets connections.

No example or OSS source code available ?


Reply via email to