Re: [DNSOP] New Version Notification for draft-hoffman-dns-terminology-00.txt

[Andrew Sullivan <[e-mail]>]

I really like this approach.  Thanks!

-- 
Andrew Sullivan 
Please excuse my clumbsy thums. 

> On Nov 30, 2014, at 9:02, Tony Finch <[e-mail]> wrote:
> 
> Andrew Sullivan <[e-mail]> wrote:
>>> On Sat, Nov 29, 2014 at 12:06:28PM +1100, Mark Andrews wrote:
>>> 
>>> A iterative resolver make RD=0 queries.
>>> A resursive resolver make RD=1 queries.
>> 
>> The problem with that definition, though I guess it is technically
>> most accurate, is that it doesn't help anyone who doesn't already
>> understand the difference.  After all, the second of these only says,
>> "a recursive resolver says that it wants recursion", which is hardly
>> elicidating for the uninitiated.
>> 
>> I think Tony's proposal is on the right track (certainly better than
>> the quick excerpt we did from STD 13, which is at least as confusing).
>> But I think Paul put his finger on the problem with that text: there's
>> no way to know in advance what sort of server you're querying.
> 
> I think it is worth noting that RFC 1034 talks about the resolution
> process, i.e. recursive resolution and iterative resolution - it does not
> talk about iterative resolvers or recursive resolvers. Which sort-of makes
> sense since whether a resolver is being iterative or recursive is to a
> large extent a matter of configuration - as Mark said BIND can do both,
> and this is also true for Unbound, ldns, getdns-api, etc. The exception is
> pure stub resolvers (e.g. the libc resolver or adns) which can only
> perform recursive queries.
> 
> So I think to clarify the terminology, first describe the iterative and
> recursive resolution processes, and mention the expected values of the
> RD,RA,AA bits. Then you can describe the distinctions between recursive
> and iterative clients, and between recursive and authoritative servers, in
> terms of the roles they play in the different resolution processes.
> 
> I would say a full resolver is one that contains a cache and is capable of
> iterative resolution [citation needed]. A stub is (traditionally)
> cacheless and recursive only. There are intermediate cases - a lot of
> stubby resolvers have caches.
> 
> Tony.
> -- 
> f.anthony.n.finch  <[e-mail]>  http://dotat.at/
> Rockall: Southwesterly 4, increasing 5 to 7, occasionally gale 8 in northwest,
> veering northwesterly 4 or 5 later. Moderate or rough, occasionally very rough
> later in northwest. Occasional rain. Good, occasionally poor.