On Wed, Dec 20, 2006 at 05:39:45PM +0000,
 Paul Vixie <[EMAIL PROTECTED]> wrote 
 a message of 2412 lines which said:

> my read on this document is that it describes something more
> expansive than the ISC tool mentioned above.

Yes, because it includes things like ACL (possibly), server reloading
or views. But one of the purposes of the discussion is to define the
scope of the future protocol: constrained (just a few things like zone
provisioning) or large (everything related to DNS server management)?

> it's more like "junoscript or netconf for DNS authorities",

IMHO, no, we should not try to define a standard way to describe
everything which goes in named.conf or nsd.conf. We should focus on a
few things but it is not obvious what these things are.

[Is it just me or is there a lot of NIH in Netconf, such a filtering
language for XML, instead of XPath and a new RPC protocol when XML-RPC
or SOAP, not mentioning REST, already exist?]

> the attached paper was published last year but i've not got around
> to writing an I-D for it nor releasing the software that implements
> it.

The way I view your "metazone" proposal is as a candidate for the
protocol whose requirments are expressed in
draft-regnauld-ns-communication-00.txt. My vision was that
draft-regnauld-ns-communication-00.txt would become a "Requirments"
RFC and then we could evaluate several candidate protocols.

> my big worries are that it's either too BIND-specific

At least, the vocabulary is, with the names of the zones being
named.conf directives :-)

Part of the challenge for the future protocol will be to define a
standard vocabulary. "Stub" and "forward" zones are another good
example.


_______________________________________________
DNSOP mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/dnsop

Reply via email to