[Prev][Next][Index][Thread]

Re: Survey of the mail server software in a top-level domain

At 02:35 PM 8/1/97 +0200, Stephane Bortzmeyer wrote:
>
>Summer being the good time for small experiments, I ran a test
>on a set (randomly choosen) of the mail servers in one top-level domain.

Most interesting. I think Dan must have a similar program which he runs on 
sporadic occassions. As someone else has mentioned, the help response is 
also a useful differentiator.

>We get the following:
>
>1078 different servers tested (2192/8763 domains):
>sendmail:   533 (49 %)

>lotus:   15 (1 %)
>probably qmail:   11 (1 %)

>zmailer:   2 (0 %)
>smail:   2 (0 %)
>pp:   2 (0 %)


It also shows, in your sample leastwise, that qmail is already ahead of 
smail, pp and zmailer which have been the most obvious alternatives to 
sendmail for quite a few years.

So, assuming one million flies can't be wrong (sorry an old aussie 
expression) then qmail is doing rather well in it's short life.

But I agree that:

>which seems to indicate that the real sendmail-killer is not qmail but 
>the various software running on Windows-NT, although they are still less 
>common than sendmail+Unix, even all together.

The other mailer I left un-snipped was Lotus. I've have the interesting task 
of bringing a Lotus SMTP server online recently. It re-confirms our worst 
fears that technical and implementation competance have nothing to do with 
market success.


Regards.