[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: failure detection
Hi Erik,
On Fri, 26 Aug 2005, Erik Nordmark wrote:
I had one minor comment on the above.
It might very well be a useful optimization for routers to be able
to express to hosts which prefixes are currently not working for
going outside the site.
Yeah, probably. This wouldn't even be shim6 specific.
But I think it would be a mistake to overload the
preferred/deprecated notion for this purpose; if we want to do this
we can just as well add a "temporarily unavailable" bit to the
prefix information options, and that way avoid confusing any hosts
that are not shim6 aware.
You're the expert. :)
One thing though: from the POV of the router which wants to change
RA's due to link-failure, it has no idea of how temporary the failure
is. So maybe call it 'failed' instead? But that implies continuing to
send RAs for the prefix.
I was envisaging that the router would simply stop sending RAs, and
once the preferred time ran out, hosts would use other prefixes (if
any were available).
It remains to be seen how hard it would be to provide this for a
site with multiple routers. We have the router renumbering RFC,
which in theory could be used to carry this from a border router to
other routers in the site, but nobody is implementing this protocol
AFAIK.
Maybe because no one is interested in multi-prefix IPv6 sites until
solutions like shim6 are available?
So it's not clear to me whether such a mechanism would be limited
to the single link site where the RAs come directly from the border
router.
Hmm, I guess yes. The multi-router case and protocol for distributing
external link-availibility seems a general one.
I guess an implementation could offer a facility to 'pin' RAs to a
specific prefix announced in IGP. IGP withdraws route, the routers RA
code either stops sending RAs for that prefix or uses the 'failed'
bit?
Erik
regards,
--
Paul Jakma paul@clubi.ie paul@jakma.org Key ID: 64A2FF6A
Fortune:
Why are you so hard to ignore?