ietf-mxcomp
[Top] [All Lists]

Re: Semantics: per user policy

2004-05-08 08:31:48


I'm sorry, I thought you we're making a point other than the "roaming user"
issue.

I think he was, in his original message.  Let's take my address,
wfenner(_at_)att(_dot_)com(_dot_)  Out of 66971 email addresses @att.com, there 
are 135
incoming mail servers (some of which are departmental servers which have
further subdivisions into incoming servers).  There are also external
forwardings to 390 other domains.

Right now, there is no particular policy about outgoing mail.  Some
large fraction probably use the incoming mail server that is known
to the @att.com forwarder.  A large fraction of the rest probably use
the second-level incoming mail server.  There's unquestionably a small
collection of random other uses - departments with outgoing servers
that are not their incoming servers, or just individuals with their
own outgoing servers (e.g., when I send mail from wfenner(_at_)att(_dot_)com,
I use my local workstation as my outgoing mail server.)

I think Roy's point was that an organization of this size has at least
135 entities to be concerned about getting to agree, and if even one
of the 135 balks, it's tough to have a unified policy for the domain.
That's even ignoring the 390 external domains that legitimately contain
@att.com addresses, each of which might need its own special handling.

  Bill

[data retrieved from the employee directory LDAP database; this is either
 itself the source of the data that the att.com incoming servers use to
 forward mail or is fed from the same source.]