ANY queries and Recursion

[Kevin Darcy]

Eric A. Hall wrote:

> > The value is that not many clients, if any, really need *full* answers
> > to QTYPE=* queries, and it's a burden on the server to provide them.
>
> There are clients which do need ALL though, and they are asking for ALL
> and are expecting ALL in the answer.

For instance?

And how many of those clients are inherently *incapable* of finding an
authoritative answer for themselves?

Stub resolvers are a necessary evil, IMO. I'd rather see *all* resolvers do
caching, follow referrals, etc. assuming they have the necessary connectivity
(obviously recursion has its place in firewalled environments). Now you want
to make matters *worse* by putting the onus on a server to search out every
possible RRset in response to every QTYPE=* query?!?! Sorry, I just can't go
along with that plan; it goes 180 degrees from where I think we should be
heading. You seem to want to dumb down endpoint resolvers even more than they
already are, whereas I think they should take on a *higher* level of
responsibility for their name-resolution functions. Spread the load, I say.


- Kevin