Missing ROUTE_REJECTION capsule about draft-ietf-masque-connect-ip HOT 7 CLOSED

Oh, is that what the rejection capsule was for? I guess I'd just rather have the route advertisement itself provide a list of exceptions, rather than tying it to a separate capsule?

from draft-ietf-masque-connect-ip.

Yes, that would also work. This is just a matter of encoding.

from draft-ietf-masque-connect-ip.

The nice thing about the rejection capsule is it allows you to remove announcements that were previous present without having to re-send a full routing table. I would prefer keeping it, as it allows for more efficient transmission of information. (I disagree it's just a matter of encoding.)

from draft-ietf-masque-connect-ip.

Actually sending a full routing table seems like the safer way to go. And given you don’t do that very often, I don’t think we need to worry about overhead.

from draft-ietf-masque-connect-ip.

I will disagree with my co-author. I think the basic case for what in most cases are the equivalent to the default route for a host on a LAN segment would be good to have in a reasonable compact form. I think the question to ask here is if we in general need a mechanism to specify an IP prefix with one or more carved out sub-prefixes. I that requirement may apply to both advertising and rejecting/removing a prefix. In the removing maybe it is possible to specify the whole prefix without carve out sub-prefixes.

from draft-ietf-masque-connect-ip.

Maybe call with ROUTE_WITHDRAW instead of reject

from draft-ietf-masque-connect-ip.

ROUTE_WITHDRAW sounds good to me.

from draft-ietf-masque-connect-ip.