>>Actually, if you are unfortunate enough to have a list of addresses sorted
>>by the right side of the @, qmail can be a big loser here. ...
>somedomain is poorly configured. Should qmail assume all sites are
>poorly configured? Should properly configured sites suffer because
>some sites are poorly run?
This is a topic I've been thinking about for a while. I want my MTA
to open as many connections to a remote site as the remote can handle.
If we're lucky, the remote is well configured and will reject
connections to tell us that it's busy, but as we all know, most MTAs
tend to accept them all and fall over.
Since qmail already keeps a retry interval for each remote IP it tries
to contact, how hard would it be also to keep some estimate of the
remote load, perhaps the time from accepting the connection to sending
the initial banner? Then it could limit the number of simultaneous
connections to slow hosts, for some definition of slow.
On a site with a lot of outgoing mail, I'd think this could improve
overall outgoing mail throughput, since it prevents qmail from doing a
DOS on itself by opening all of its outgoing connections to hosts that
are terminally slow. Instead, it favors deliveries to hosts that can
accept mail quickly and gets them out of the queue.
Yeah, none of this should be necessary, but there are a lot of
features in a robust MTA that wouldn't be necessary if the rest of the
world were better behaved.
--
John R. Levine, IECC, POB 727, Trumansburg NY 14886 +1 607 387 6869
[EMAIL PROTECTED], Village Trustee and Sewer Commissioner, http://iecc.com/johnl,
Member, Provisional board, Coalition Against Unsolicited Commercial E-mail
