[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: IOCP / C10K



> On 3/7/07, Toby Douglass <toby.douglass@xxxxxxxxxxxxxx> wrote:
>> I've written a C10K solution for Windows, libiocp.
>> ...
>> It is useful in its current form?
>
> this is useful for libevent.  unfortunately i am not volunteering a patch
> :)

They know about it.  I've removed them from the CC:.  The problem is that
the way libiocp is written is, as I understand it, not particularly
compatable with the way libevent is written.  Merging the two is not an
obvious proposition.  How focused is the Tor code on libevent?  is it only
used for sockets?

> note that even with overlapped i/o or completion ports many xp/2k
> kernels max out buffer resources at 5k or so open sockets (sometimes
> much less (or server variants much more)).

I have found that I can have far more idle sockets than active sockets. 
I've not experimented to see if an active socket gone idle releases
resources.