On Sun, Sep 25, 2011 at 12:19 PM, Paul Hoffman <paul.hoff...@vpnc.org>wrote: > On Sep 25, 2011, at 11:26 AM, xun...@isi.edu wrote: > > > We have a draft of our proposal with rationale at > > http://www.isi.edu/~xunfan/research/draft-anycast-diagnostics.txt > > Instead of "ns-diagnostics", using "_ns-diagnostics" would be much better. > The "_ at the beginning of a label is not a hostname" meme seems to be > accepted by nearly everyone in the DNS world. Sure, thanks for the advise. We can add "_" for "instance_id" and "node_id" also, "_instance_id._ns-diagnostics.$ORIGIN", "_node_id._ns-diagnostics.$ORIGIN" > > You do not define the difference between an "instance" and a "node", and > RFC 4892 doesn't define "node". Either clearly differentiate them here, or > maybe just get rid of "node". Yes, the is not the first time we have this terminology problem. "node" is defined in RFC 4786 while RFC 4892 use "instance". We only have a slight explanation of the two in Section 1 the first goal, "The mechanism should be able to identify both specific anycast instances (a specific computer running anycast), and also anycast nodes (defined in RFC 4786 [RFC4786] as the topology location where multiple instances may run)" but we should definitely give a clear differentiation at the beginning. Thanks! > > --Paul Hoffman > > _______________________________________________ > DNSOP mailing list > DNSOP@ietf.org > https://www.ietf.org/mailman/listinfo/dnsop >
On Sun, Sep 25, 2011 at 12:19 PM, Paul Hoffman <paul.hoff...@vpnc.org> wrote:
On Sep 25, 2011, at 11:26 AM, xun...@isi.edu wrote:
> We have a draft of our proposal with rationale at
> http://www.isi.edu/~xunfan/research/draft-anycast-diagnostics.txt
Instead of "ns-diagnostics", using "_ns-diagnostics" would be much better. The "_ at the beginning of a label is not a hostname" meme seems to be accepted by nearly everyone in the DNS world.
You do not define the difference between an "instance" and a "node", and RFC 4892 doesn't define "node". Either clearly differentiate them here, or maybe just get rid of "node".
--Paul Hoffman
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www.ietf.org/mailman/listinfo/dnsop
_______________________________________________ DNSOP mailing list DNSOP@ietf.org https://www.ietf.org/mailman/listinfo/dnsop
