[dns-operations] IPv6 & IPv4 addresses

[Edward Lewis]

At 14:13 +0000 3/17/11, Simon Munton wrote:

>So my question is - is there a case for always returning *both* A and AAAA
>records in response to an A (or AAAA) query - can anybody see any harm in
>doing this? (other than a slight increase in response packet size).

The connectivity of the IPv4 internet is more mature than the IPv6 
internet.   (I'm not judging, I'm just saying...)  Because of this, 
some content providers discourage the use of IPv6 for the general 
population.

Here's a better description:

http://tools.ietf.org/html/draft-livingood-dns-whitelisting-implications-01

>Or is there a case for creating a new, virtual RR, called (say) ADDR so a
>query for ADDR would mean "give me both A & AAAA" ?

For backwards compatibility you need to maintain the existing 
behavior waiting until this record becomes ubiquitous.  There's a 
catch-22 here, the reason this kind of transition gambit won't solve 
the problems is that we can't move off the old broken software 
everywhere.  If we could move off "the old broken software 
everywhere", then we wouldn't have (v4->v6) transition issue to begin 
with and the gambit wouldn't be needed.

If you say "let's try anyway" - it would be easier to just say to 
anyone that would use the ADDR record to just ask for A and AAAA 
records.  It would take a small amount of more code and save all of 
the standards work needed.

(FWIW, this idea has been floated before.  I"m just summarizing what 
ultimately shot if down.  Decent idea to have, but ...)
-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Edward Lewis
NeuStar                    You can leave a voice message at +1-571-434-5468

Me to infant son: "Waah! Waah! Is that all you can say?  Waah?"
Son: "Waah!"