[dns-operations] RFC 6382: Unique Per-Node Origin ASNs for Globally Anycasted Services

[Danny McPherson]

On Oct 25, 2011, at 3:45 AM, Stephane Bortzmeyer wrote:

> Relevant for the DNS, since the biggest user of the anycast is
> probably the DNS. Today, the anycast operators have different BGP
> practices. Is it a good idea to have one different AS per site? The
> RFC says so:
> 
> http://www.rfc-editor.org/rfc/rfc6382.txt
> 
> See also, for a counter-point :
> 
> https://www.isc.org/community/blog/201109/origin-asn-anycasted-services
> 
> Interesting for debugging anycast, anyway.

For the record, this RFC publication was essentially done (i.e., completed 
AUTH-48) at the time the ISC blog post appeared and it was discussed point 
by point on the mailing list (and in a subsequent NANOG thread on operational 
issues just days later).  The _authors asked the RFC Editor to hold publication 
and discussed all of your comments with the IESG and OPS ADs, and after they 
considered, they asked that the hold be removed and the document published as 
is.  

It should be clear that this publication in no way precludes alternatives, 
assuming the WG/community finds merit in them.

For more information on the discussion regarding the ISC blog post within 
GROW and a related NANOG thread see these threads:

<http://www.ietf.org/mail-archive/web/grow/current/msg02044.html>

and

<http://mailman.nanog.org/pipermail/nanog/2011-October/040737.html>

-danny