On Wed, 07 May 1997 19:12:52 -0400 Brad Knowles <[log in to unmask]> said:
> I've discovered over the years that it's not recipient deliveries
>that are important, since large numbers of deliveries typically get sent
>to the same MXes (so what you would need to count instead would be
>number of separate envelopes delivered).
Well, I've discovered over the years that sendmail people invariably
think in terms of the things that limit sendmail's performance. Other
systems have other strengths and weaknesses, and different limitations.
> We received over five million messages yesterday (and rejected
>another seven million delivery attempts), to a total of almost 23
>million recipients.
So, we do have more than 1/10th of your volume.
>The Internet side is typically about half the total volume of the AOL
>mail system as a whole.
Internal mail from AOL user 1 to AOL user 2 doesn't go through SMTP and I
don't see how it is relevant to this discussion, other than conveniently
providing the 23M of deliveries you were missing :-)
> Now, how many million messages did you receive yesterday?
Receive, not that many. There are the bounces of course, but like any
other large mailing list shop, we receive a lot less than we send. I
imagine AOL is the opposite and receives a lot more than it sends.
Anyway, I think what you really want to know is the number of SMTP
transactions that we've made, regardless of the recipient count, right? I
can get exact numbers if necessary, but it's usually about half the
recipient count (lots of small to mid-size lists with mostly one
subscriber per host), so about 2 million and change, vs your 5 million
SMTP transactions. Ironically, you had a much more impressive workload
when we were talking recipients :-)
>> As a matter of fact, yes, I do think I could do better than your
>> sendmail-based setup!
>
> Then put your money where your mouth is. Give up your day job and
>come over here to learn how the big boys play the game.
I have no intention of either giving up my day job, moving to the US or
joining AOL, nor do I see any reason why this would be necessary in order
to accomplish the stated goals. Nevertheless, I was making a serious
business proposal. I am not in the habit of saying "I can do this!" and
then explaining that I was really just kidding to try and impress people,
and could everyone please forgive me. I was not kidding at all and while
you obviously think that no one outside of AOL can possibly understand
the challenges that AOL is facing, I have heard that tune before and
there is nothing I like more than a technical challenge, so please do
send me your technical requirements and we will follow up on them. And
please don't tell me that this is a waste of time because I can't begin
to imagine how incredibly momentous your requirements are because I'm
just a little boy lost in the woods, this is simply not going to convince
me. In 1987 we had a mainframe which very nearly had a terabyte of data,
and that was a LOT of data back then. Yes, this and other factors did
create "in kind" problems as you said, but we sat down and worked on them
until they were solved. You just didn't get responsibilities in the
mainframe world unless you could sit down and solve "big" problems
without letting the numbers impress you out of your wits. It is by
applying "big" problem skills to more affordable hardware (and not saying
"It can't be done! It's way too much traffic!") that products like LSMTP
are developed. I see that Matt Korn is still your VP of Operations, so it
looks like I may be preaching to the choir :-) Anyway, just take a piece
of desktop, write down all the numbers, however big, add all your special
requirements, and let us work on them. Maybe we won't be able to meet
your requirements right now, but software is made to be improved.
Eric
|