Re: Qmail metrics

• To: djb-qmail@xxxxxxxxxxxxxxxxxxxx
• Subject: Re: Qmail metrics
• From: Dave Sill <de5@xxxxxxxxxxxxxxxxx>
• Date: 27 Aug 1996 19:20:53 -0000
• Delivered-To: djb-qmail@xxxxxxxxxxxxxxxxxxxx
• In-Reply-To: <19960827183303.1493.qmail@xxxxxxxxxxxxxxxxxxxx>
• Organization: Oak Ridge National Lab, Oak Ridge, Tenn., USA
• References: <19960827183303.1493.qmail@xxxxxxxxxxxxxxxxxxxx>

I've never heard of any MTA benchmarks, and being the comp.benchmarks
FAQ author I'm in a pretty good position to hear about them. Such a
tool would be an excellent way to determine where qmail is faster and
by how much. It'd also be nice to be able to tell what impact a new
release or system reconfiguration had on mail precessing speed.

What kinds of things would one want to measure, and how would one go
about it? Off the top of my head I'd like:

    messages/sec and MB/sec delivered to local user
        for various message sizes
        for various numbers of recipients
        via SMTP and qmail-inject or sendmail
        to file, maildir, bit bucket (/dev/null), null program (/bin/true)
    messages,MB/sec delivered to remote user
        would have to deliver to multiple remote systems, and would
            want to know which MTA remotes were using
        for various message sizes
        for various numbers of recipients
        via SMTP and qmail-inject or sendmail
        to file, maildir, bit bucket (/dev/null), null program (/bin/true)
    bounces/sec to nonexistent local users

Any comments or other ideas? Respond directly to me, please.

-Dave

• Re: Qmail metrics
• From: djb@xxxxxxxxxxxxxxxxxxxx (D. J. Bernstein)

• Prev: Re: Qmail metrics
• Next: Re: Qmail looks great!
• Index(es):
  • Main
  • Thread