On 26-Jul-23 01:07, Michael Richardson wrote:
Brian E Carpenter <brian.e.carpen...@gmail.com> wrote: > Yes, but you can map them in CBOR just as > draft-eckert-anima-grasp-dnssd already describes. (Think JSON but code > CBOR.) My only real concern is how to extend the objective for the > AN_join_registrar. It seems lame to use plain text when a JSON-style > map would be much nicer for the programmer. My prototype code for > grasp-dnssd has to parse DNS records to change them into Python > maps. But once that's done, sending or receiving them as CBOR is > trivial. so far we are sending multiple objective values: https://www.ietf.org/archive/id/draft-ietf-anima-constrained-join-proxy-14.html#name-grasp-discovery and https://www.ietf.org/archive/id/draft-ietf-anima-constrained-voucher-21.html#name-grasp-discovery [M_FLOOD, 51840231, h'fda379a6f6ee00000200000064000001', 180000, [["AN_join_registrar", 4, 255, ""], [O_IPv6_LOCATOR, h'fda379a6f6ee00000200000064000001', IPPROTO_TCP, 8443], ["AN_join_registrar", 4, 255, "BRSKI_JP"], [O_IPv6_LOCATOR, h'fda379a6f6ee00000200000064000001', IPPROTO_UDP, 5684], ["AN_join_registrar", 4, 255, "BRSKI_RJP"], [O_IPv6_LOCATOR, h'fda379a6f6ee00000200000064000001', IPPROTO_UDP, 5685]]] I have resisted suggestions that we put an array for the objective-value, and also that it have a string that needs to be parsed like "mode=prm,foo=1,bar=2"... This is because I think that the protocol and port numbers, and maybe even the IP addresses will be different.
That makes sense, but it probably needs to be normatively specified, which would avoid any need to change RFC 8995, except perhaps an erratum to delete that "(list of)" wart.
(IP addresses could change not because different machines, but because in IPv6, why not, and it also makes use of containers easier) I regret we didn't write "BRSKI_EST" on the first one. I am asking now if we should have a registry for this. It could be RFC Required, as at this point, everything is in WG documents. Since we are doing this in std track documents (so they get the ultimate considerations, of IESG action, in effect), the value of the registry is that it lets people find the document that goes with the value.
Yes, makes sense. Brian
If we do want a registry, it needs to go into one of the two above documents, I think. I don't want to drag this on longer than it needs to be, because these documents have taken too long already. -- Michael Richardson <mcr+i...@sandelman.ca> . o O ( IPv6 IøT consulting ) Sandelman Software Works Inc, Ottawa and Worldwide
_______________________________________________ Anima mailing list Anima@ietf.org https://www.ietf.org/mailman/listinfo/anima