I have looked at ttps://datatracker.ietf.org/doc/draft-hares-i2rs-bnp-eca-data-model/ and https://datatracker.ietf.org/doc/draft-hares-i2rs-pkt-eca-data-model/
In general, I am comfortable with the WG adopting these documents. I think there is one important issue that needs to be addressed, preferably in conjunction with adoption. the issue as two related parts.
The primary issue is the interaction between these policies and the operational RIB in the rotuer. I suspect I know what the assumption of the wg is, but I could not find any text in the document. It may be that I simply misssed it. I was looking for something that said "all packets which do not match any policy filter are passed on to the RIB."
Related to this, I think the text needs to note that the I2RS priority scheme does not help for conflicts between I2RS policy forwarding entries and I2RS RIB entries. The policy entries will apply first, even if the RIB entries have a higher priority. I think this is an inevitable consequence of our deliberate (and I personally think sensible) choice to not deal with indirect conflicts in the protocol. I do think the document should not this effect explicitly.
Yours, Joel _______________________________________________ i2rs mailing list [email protected] https://www.ietf.org/mailman/listinfo/i2rs
