Brad Knowles wrote:
>
> At 2:28 PM +0200 1999/10/13, Andre Oppermann wrote:
> > Nice paper! Most of the suggestions are already realized in qmail and
> > postfix.
>
> I realize that Postfix already addresses many of these issues. I
> was involved in the earlier stages of the beta testing for Postfix,
> and it (and private conversations with Wietse) figured very strongly
> in the development of the later stages of that paper.
>
> In fact, I like to be able to claim that the idea for using Van
> Jacobsen TCP-style slow-start/bounded exponential backoff for
> parallelism and queue retries was an idea that I contributed to the
> development of Postfix, and is one of the many factors that make it
> superior.
That is something qmail has done since the beginning.
> Of course, that's starting down a rather religious path
> that I don't think we want to traverse.
No.
> > The best is on page three, "handle at least tens of
> > thousands messages per day per server"... About a year ago I tested
> > qmail's performace with two old 486DX100/16MB/2GBIDE and 10BaseT
> > ethernet.
>
> And Wietse has done more with less. Again, more on that
> religious issue that I don't think we want to get started.
Well, the reason these numbers are so "low" is that qmail fsync()'s
every message to the disk. Softupdates was not available at that
time. But that's not the point. I was arguing against sendmail, not
against postfix. I consider qmail and postfix equal, one is stronger
in one direction whereas the other one is stronger in other aspects.
I for myselfs choose qmail for reasons I can tell you in private
email. I have no intend to "fight" religiously for qmail, the only
thing I do is pointing out the strong aspects of qmail/ezmlm for
big lists.
> > As long as these additional performance enhancements are not public
> > available I don't care.
>
> I'm very surprised to hear that, since there are some extensions
> they've made that I believe will materially change the face of
> delivering customized content via e-mail, and will change it to the
> point where it is virtually completely unrecognizable from where we
> are today.
>
> I'm sorry to learn that you don't want to learn about protocol
> enhancements of this kind of scale.
I would like, but you just said you would (can) not tell me about it...
--
Andre Oppermann
CEO / Geschaeftsfuehrer
Internet Business Solutions Ltd. (AG)
Hardstrasse 235, 8005 Zurich, Switzerland
Fon +41 1 277 75 75 / Fax +41 1 277 75 77
http://www.pipeline.ch [EMAIL PROTECTED]
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
