ietf-mxcomp
[Top] [All Lists]

Re: record size and wild cards

2004-08-12 12:51:27

Say you have three services, _marid, _able, and _baker, each
of which has its own TXT record.

Tnx, now I got, *.example.com would be ambiguous.  But maybe
we could get around this problem with an additional "_below":

_marid.www.example.com exists => ready, use it for _marid
_baker.www.example.com exists => ready, use it for _baker

No _marid.foo.example.com => try _marid._below.example.com
No _baker.bar.example.com => try _baker._below.example.com

Quite expensive, but no more conflicts for different services.
Could that be a solution for the "don't abuse TXT fraction" ?

You're basicallly proposing a new DNS wildcard system.  That's not a
terrible idea (there's certainly been enough criticism of the existing
system), but it's a considerably larger scale change than any we've
suggested so far and MARID isn't the right place to propose it.  In
the past, people have proposed using zone cuts as a kind of faux
wildcards, which isn't a good idea either.

Keep in mind that if marid, able, and baker are each assigned their
own DNS record type, existing DNS wildcards work fine.  There is a
solution to this problem, just not one that overloads everything into
TXT records.  Too bad a large vendor that likes to make all of its
protocols extensible didn't realize that DNS is also extensible when
they wrote their DNS proxy code.

Regards,
John Levine, johnl(_at_)iecc(_dot_)com, Primary Perpetrator of "The Internet 
for Dummies",
Information Superhighwayman wanna-be, http://www.johnlevine.com, Mayor
"More Wiener schnitzel, please", said Tom, revealingly.