[dns-operations] To A or to AAAA - was Re: Signaling client

[Mark Andrews]

In message <4D34B74E.7040507 at dougbarton.us>, Doug Barton writes:
> On 01/17/2011 13:28, Edward Lewis wrote:
> > The reason for this rambling tirade is that Doug's part right but falls
> > short when it gets to determining "address records they cannot use."
> > That's impossible to "compute".  I might have a v6 set up in my house
> > but be on a ISP that can route me to f.root-server.net/AAAA but not
> > m.root-servers.net/AAAA (I've seen this).  An application can't possibly
> > know a priori to ask for f's AAAA and not m's AAAA. Trying both will
> > tell me, but not until I try.
> >
> > Applications need to be more aggressive and rugged in trying to reach
> > whatever they want off the 'net.
> 
> I agree with your thorough response, mine was designed more to indicate 
> that the right layer for the solution is _not_ the DNS.
> 
> That said, while I do agree in general that applications/OS' should have 
> better heuristics in regard to connectivity I think that it should at 
> least be possible for them to be smart enough to know what transports 
> they definitely cannot use, and avoid breaking themselves by asking for 
> them.  :)

That's trival and is done by gethostname/getaddrinfo with the right
flags already.

The problem however is that the problem machines do have working
transports for IPv6 and they do have a route the destination.  It's
just that the route they have does not work and there is no way for
a application to work out that a route it has will not work without
trying it.

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