ietf-mxcomp
[Top] [All Lists]

RE: Unified SPF: block versus factored records for HELO and MTAMA ark scopes

2004-06-24 21:55:54

Depends how big the process is.  Forking sendmail takes forever, but
qmail's server is small and snappy.  I realize that one can 
do threads on
Unix boxes, but I don't think that many MTAs actually do so.

Well if we really get into UNIX arcanae Apache actually pre-forks
the processes as a hunt group and some Unix mail servers do too.

I had an office in the AI lab near to Robert Thau trying to make 
this particular party trick work on the Unix boxen of the day. 
Not a recommended approach.

Even so you still have a process fault rather than a thread switch,
if the threads package is half way competent its performance on 
a dedicated machine should be much better - of course on a shared 
machine your 16 processes will get far more resources.


If you really want performance of course you single thread in a
unitary process and use non blocking state machines... Of course
to make that trick work you have to really understand the theory.
But I have no problem loading up a machine with several thousand
simultaneous connections using that method. Beyond that point you 
have to start writing your own tcp stack or the limitations of the 
BSD sockets libraries start to bite.

        Phill


<Prev in Thread] Current Thread [Next in Thread>
  • RE: Unified SPF: block versus factored records for HELO and MTAMA ark scopes, Hallam-Baker, Phillip <=