Hi Luuk, Thanks as always for the review. Inline: On 24/10/23 12:12, Luuk Hendriks wrote:
Hi Paolo, all, Thanks for the updates. Things that caught my eye while reading: Sec 1: Nit: in the last of the newly added bullet points, 'BGPU PDU' is the only occurence where 'BGP Update' is (in my opinion unnecessarily) abbreviated this way.
Ack, 100%
Sec 4.2.3: Type TBD4 regarding ADD-PATH: with values 0/1 representing false/true, what would it mean if this TLV is not present in the message? Note that in the implementation notes added in path-marking-tlv-00 , one of the mentioned options to implicitly signal 'no path status' is to omit the marking TLV altogether. If we opt for that, it might be nice to be consistent with it here and only attach a TBD4 of length 0 (so no value) if a route is ADD-PATH.
For me - and this is currently under specified - if a specific stateless parsing TLV is not present, one should fall back to look at the BGP Open message in the Peer Up (just like today); because this TLV is optional and it may be that - while the implementation does support TLVs - it opts to not support stateless parsing TLVs.
If we agree on the above, then we need a 3 states outcome (which is not the case for the Path Marking draft): TLV not present, fallback; if TLV is present then explicitly say whether, for example, ADD-PATH is enabled or not with 0/1 values.
Thoughts? Paolo _______________________________________________ GROW mailing list [email protected] https://www.ietf.org/mailman/listinfo/grow
