Re: well-known addresses / was DNS discovery

["Eric A. Hall" <ehall@ehsco.com>]

Mark.Andrews@isc.org wrote:

> 	It is easy to accidently introduce single points of failure
> 	into anycast solutions even though you have topologically
> 	spread your nameservers.  It much harder to do this with
> 	a non-anycast solution.

Agree completely. Anycast can work well when (1) you control the
end-points and (2) you control the netblock. The current proposal leaves
those elements under somebody else' control ('default' configs that use
hard-coded addresses which follow competing route advertisementss).

> 	WKA doesn't remove the need to have another solution to
> 	supply the search list.

I don't see that as a problem for the 'default' service since the data in
question is particular to each administrative domain, and therefore has to
be configured explicitly on a per-site basis. So for that particular data,
the admin either has to config the host anyway, or the admin should be
looking towards making a management investment in DHCP or whatever suits
their needs. But for a 'default' service that just needs to be able to
resolve basic FQDNs, being able to have resolvers locate functional
servers automagically is good enough, and is a reasonable scope.

-- 
Eric A. Hall                                        http://www.ehsco.com/
Internet Core Protocols          http://www.oreilly.com/catalog/coreprot/

#----------------------------------------------------------------------
# To unsubscribe, send a message to <dnsop-request@cafax.se>.