Hi Massimiliano, On 11 Jul 2019, at 14:43, Massimiliano Stucchi <[email protected]> wrote:
> What I don't like (or most likely don't understand the necessity of) > about the draft is the jump up 2 protocol versions. This is most likely > going to create many issues for implementations and create lots of > unnecessary code in both validators and routers. Well, what I was thinking is to allow implementations to select which mechanisms they want to support. You could simply receive the Deny and/or Allow PDUs and ignore those without trouble. But Path PDUs can't exist at the same time as Prefix PDUs, so the cache server and the router MUST be in agreement on supporting that one. So version 2 adds Deny and Allow but not Path, and version 3 also adds Path. What do you think the problems will be with more versions? We already have version 0 and 1, so in any event we'd have three versions. > I think that the general idea and direction are good, but I also think > we should discuss first about how not to bloat up the protocol, and make > something that can stand long term and can be easily implemented and > maintained. Right. > I would be happy to join the discussion and participate in this, if we > see enough interest. Great! Thanks, Iljitsch _______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
