mike <[email protected]> wrote: > On 7/4/13 7:18 AM, Michael Richardson wrote: >> >> (*) THERE IS STILL an issue of what do hosts and routers do if there >> are two ISP connected routers, and potentially two completely >> different sources of DHCPv6 information. Even in the simplest two ISP >> case, hosts can see both of these DHCPv6 servers directly. >>
> Maybe there's really a few things going on here? There is clearly some
> ISP specific information that ought to be relayed to the host. But
> there's also "configurator of last resort" insofar as since there
> generally isn't much clue in the routers, the ISP's fill the void by
> providing that information. Sometimes it's benign, sometimes it's in
> the ISP's self interest. And if we have two ISP's due to multihoming,
> there's going to be conflict if it's of the self-interest kind for the
> ISP's.
It's also where a clueful user and/or a very-smart router (Apple ones come to
mind) tell hosts about various services that might be present that the hosts
care about. NTP servers, WINS servers, ... well the v4 list is:
http://www.iana.org/assignments/bootp-dhcp-parameters/bootp-dhcp-parameters.xhtml#options
> As far as I can tell now, it would just be a race in the multihoming
> situation, right? Pretty much whichever DHCP server responded first
> (or their relay) would win which could cause some inconsistency for
I think so.
This is why I think that the router configuration protocol isn't about
talking to hosts or about replacing DHCP, it's about DHCP servers getting
together to be consistent in what they tell hosts.
> configuration that is "shared". DNS, for example, might give different
> results if, say, one of my upstream providers answered queries
> differently than another because one has, say, a service to do split
> horizon DNS to fake up names for my home and the other doesn't. Or if
> you hate that example, there have got to be dozens of others in the
> kitchen sink of DHCP options.
DNSSEC to the rescue :-)
> I'm not exactly sure what ought to be done about that because each ISP
> is incented to be the ultimate truth of configuration information. It
> seems to me that the only arbitrator for such problems is likely to be
> a human being at home dealing with this.
Well, I think that an ISP that doesn't give the home user access to the
router doesn't deserve any money, but I understand not everyone cares.
--
] Never tell me the odds! | ipv6 mesh networks [
] Michael Richardson, Sandelman Software Works | network architect [
] [email protected] http://www.sandelman.ca/ | ruby on rails [
--
Michael Richardson <[email protected]>, Sandelman Software Works
pgpvWQRMV6ZpP.pgp
Description: PGP signature
_______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
