Several sites have measured the switch from sendmail to qmail. In every
case, sendmail used slightly more bandwidth.
It turns out that qmail's savings in DNS lookups, SMTP text, et al. are
more important than sendmail's savings in multiple RCPTs. I wouldn't be
surprised if exim uses more bandwidth than qmail, for the same reasons,
but nobody has done the necessary measurements.
Nigel Metheringham writes:
> Dan maintains that bandwidth conservation is pointless - bandwidth is
> cheap -
When did I ever say any such thing? Reference and exact quote, please.
What I keep telling people is to profile, not speculate. At most sites,
mail is below 5% of overall Internet traffic, mailing-list messages are
below 1%, and the maximum possible benefit from multiple RCPTs is below
1/2000. Measure your costs! Fanatical microoptimization doesn't produce
good results.
If you find that mail traffic is a serious cost, you shouldn't be using
SMTP. For example, qmail's QMQP can blast a 1000-recipient mailing list
through a 28.8 modem in 10 seconds. If you're doing the same thing with
sendmail---typically 1000 seconds and 50 times as much traffic---why do
you expect anyone to believe that you care about your bandwidth?
> we abuse systems that have had temporary outages less than a qmail system
Is there any evidence to support that claim? Unlike some mailers, qmail
doesn't opportunistically bombard a host that reappears on the network.
It waits patiently for the natural retry time for each message.
---Dan
50000 new aliases in 6 seconds.
http://pobox.com/~djb/fastforward.html
--
*** Exim information can be found at
http://www.exim.org/ ***
