🚀 go-pugleaf

barr@pop.psu.edu (David Barr) writes:

> In article <1992Dec19.163210.12770@becker.GTS.ORG> uucpadm@becker.gts.org wri
> >	It seems logical for sites which are Internet/UUCP
> >	gateways to declare MX records for all the systems
> >	that they publish in their pathalias map entries -
> >
> >
> >		sitex.uucp.	MX	10	mail.mysys.mydom.
> >		sitey.uucp.	MX	10	mail.mysys.mydom.
> >		...
> >
> >	I know we don't do this at my workplace, but there
> >	seems no obvious reason not to do so.
> 
> 	Most mailers are configured by default to not do DNS lookups on
> certain specified domains.  (Most notably .UUCP, .BITNET, and the
> now-gone .CSNET)  This is to save undue wear-and tear on the root
> nameservers.

Yes, but... I've come across more than a few mailers that (incorrectly)
attempt to do DNS lookup on .UUCP sites.  I've had several direct exper-
iences with failed mail in exactly this scenario.

> 	Even if this were not so, with DNS there's no way for one site
> to just "declare MX records" for .UUCP and have them be seen by just
> anyone.  _One_ site has to be declared as "authoratative" for entire
> .UUCP pseudo-domain in order for anything to work consistently.  How is
> my mailer going to know what nameserver to look at to resolve the MX
> for "becker.UUCP"?

Hmmm... some site would need to advertise itself as a top-level server
for .UUCP?  Not likely, not for free?  Or is it really all that big a
deal?  Has any site actually ~tried~ it, to show what the results would
be?  Could the MX record on a given site still be made to point to the
true best path to the UUCP-connected site?

For example, suppose UUNET declared itself the top-level server (and,
thus, only server, as .UUCP is not sub-domain-ized <ick>) for .UUCP.
If the MX record for willard.UUCP were to indicate a path of
gatech.edu!vdbsan!willard!%s, or something similar, would that not still
route mail in the exact same way that UUNET would route mail otherwise?

Mail would not need to be sent to UUNET in order to see that the route
to willard.UUCP would best be through the indicated path.  So, a tighter
link between MX/pathalias would not be necessarily a bad thing, unless
the resulting load on the system acting as the nameserver were such that
it would in fact be inpractical to implement.

What have I overlooked?

> 
> --Dave
> -- 
> System Administrator, Population Research Institute    barr@pop.psu.edu
> 	What if there was no such thing as a hypothetical question?


-- 
dawson@willard.UUCP (Willard Dawson)
gatech!vdbsan!willard!dawson
emory!uumind!willard!dawson
Willard's House BBS, Atlanta, GA -- +1 (404) 664 8814