non-24 bit subnets

[Mark Andrews]

In message <4CACDF3C.9040706 at chem.umass.edu>, Alex McKenzie writes:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> 
> 
> Jay Ford wrote:
> > On Wed, 6 Oct 2010, Alex McKenzie wrote:
> >> Out of curiosity:  what if it's a /16 or /8 network?  Do those also get
> >> built as 24 bit files, or can they be built differently?  I seem to
> >> recall seeing an option for a reverse lookup file with hosts declared as:
> >>
> >> x.y    PTR    host.domain.tld.
> >>
> >> Does that work, or was that an old format that's been deprecated, or
> >> would it never have worked?
> > 
> > Sure, that works
> > 
> > For the /16 case, define the zone like b.a.in-addr.arpa & define records
> > like
> > "d.c PTR name." for address a.b.c.d.
> > 
> > For the /8 case, define the zone like a.in-addr.arpa & define records like
> > "d.c.b PTR name." for address a.b.c.d.
> > 
> > Note the order of the address components in the zone file, with least
> > significant furthest left.
> 
> Got it.  So basically bind can cope with a subnet that falls on an octet
> boundary, but not inside an octet.  That's unfortunate for my purposes,
> but not unreasonable.

A better description is "the DNS can cope ....".  Basically it is
a well known mapping from IP addresses in to the IN-ADDR.ARPA
namespace.  This mapping has no knowledge of the subnet boundaries.

Mark
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org