I understand "cache-only" or "caching-only" DNS server as being, strictly
speaking, one which loads *no* authoritative data. Typically, this is a
resolver which populates its cache by initially priming with some "root hints"
configuration, and then walking down the namespace hierarchy via iterative
queries for everything else, but I suppose, arguably, a forwarder could also
qualify as a "cache-only" or "caching-only" DNS server too. Sometimes I've
heard the term -- despite being absolute -- being bent to include nameservers
which only load a *minimal*, "convenience" set of authoritative zones, e.g. for
"localhost", "1.0.0.127.in-addr.arpa", zones of that nature, each of which are
usually intended to provide resolution for a single, non-globally-unique name.
A "stub zone" is one in which the source of zone information -- typically in
the form of IP addresses from which to fetch the data -- is defined statically
within the nameserver configuration, but *not* involving full replication of
the zone *or* recursive resolution towards that source, hence distinguishing it
from being a slave for the zone, or engaging in any form of "forwarding",
respectively. Because there is no full replication, whatever contents of the
zone exist on the instance configured as a "stub" for it, are not considered
"authoritative" and the instance does not respond authoritatively for the zone.
- Kevin
-----Original Message-----
From: DNSOP [mailto:[email protected]] On Behalf Of Paul Hoffman
Sent: Tuesday, February 24, 2015 1:53 PM
To: Declan Ma
Cc: IETF DNSOP WG
Subject: Re: [DNSOP] New version of the DNS terminology draft
On Jan 20, 2015, at 7:56 AM, Declan Ma <[email protected]> wrote:
> As for 'DNS Servers', I think we should set aside space for 'Cache-only DNS
> Server' which is pervasive in all kinds of DNS document.
Can you clarify what you think a "cache-only DNS server" is? I'm not seeing how
a server can be cache-only without it also doing recursive queries.
> And as in 'Zones', you mentioned 'Origin'. So, I suggest adding a paragraph
> to describe 'Default TTL', which is represented as $TTL in zone file.
We can add that.
> Still, 'Stub Zone' is yet another common expression in DNS context. We may
> put it into this document.
I don't find "stub zone" mentioned in any RFC, so I can't really imagine what
that is. Please clarify.
--Paul Hoffman
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop