From: Paweł Leśniak (no email)
Date: Fri Apr 25 2008 - 04:18:31 EDT
mouss pisze:
> most backscatter comes from servers which accept mail and only later
> find that the recipient does not exist, so they bounce. if they
> validated recipients during the smtp transaction, there would be no
> bounce.
OK. And we have two problems here. One - other servers have "corrupted"
configurations (sometimes I believe one wants this kind of behavior).
And the second one - if MX of domain used at reply-to address has no SPF
record I have to use lots more resources to do body tests and find if
bounced message was sent through my server originally. I don't know if
there's any other good way to find out if this is backscatter or not
without expensive body tests. And so I have to accept message. And than
what? Should I bounce it ;)
And just like Wietse wrote - most backscatter emails comes from
nonexistent addresses.
>>> the requirement does not mean that you must accept any mail to
>>> postmaster. the requirement is to make sure you won't miss important
>>> mail to postmaster and abuse.
>> And how do I know id mail is important or not?
>
> the sender knows! if his mail is important, he must setup his email
> server correctly.
So once sender's MX must obey to standards, and the other time
(backscatter) doesn't really have to.
>> OK, but still I think that SAV with caching is not hurting badly. I'm
>> not sending millions emails daily.
> you do SAV on mail you _receive_, not on mail you _send_. so the
> number of mail you send is irrelevant. you do not control how much
> mail you receive.
>
Sorry, I didn't mean sending actually. I was thinking about how many
email goes through my server.
Pawel Lesniak
|
|
|