Hi Donald, Thanks for your reply.
On Apr 7, 2021, at 9:22 PM, Donald Eastlake <d3e...@gmail.com<mailto:d3e...@gmail.com>> wrote: ... 2. Section 2.3: EVPN Network OAM mechanisms MUST provide in-band monitoring capabilities. As such, OAM messages MUST be encoded so that they exhibit identical entropy characteristics to data traffic in order that they share the same fate. It’s not obvious to me what you mean by “identical entropy characteristics to data traffic”. Surely, different flows may have different entropy characteristics, so, *which* data traffic? Similarly, with which data traffic are you saying the OAM messages must share fate? Well, it depends on what is meant by "characteristics". I believe the intent is to say that with a sufficient quantity of OAM messages, all paths used by actual data will be exercised. Your explanation (“all paths… will be exercised”) seems inconsistent with §3.1.1.1: - all paths. For MPLS/IP networks with ECMP, monitoring of all unicast paths between MEPs (on non-adjacent nodes) may not be possible, since the per-hop ECMP hashing behavior may yield situations where it is impossible for a MEP to pick flow entropy characteristics that result in exercising the exhaustive set of ECMP paths. Monitoring of all ECMP paths between MEPs (on non- adjacent nodes) is not a requirement for EVPN OAM. Even if the above somehow isn't the issue it appears to be, I think it’s evident that the §2.3 text in question isn’t unambiguous enough to have a MUST applied to it, I think this needs to be rectified in some way. I’m going to switch my ballot to a DISCUSS until we’ve resolved this. (Now that I’m looking at the §2.3 text again, “identical” bugs me too, since identity is a much stronger requirement than, say, similarity. But since I suspect the text will be rewritten anyway it’s not worth dwelling on.) Thanks, —John
_______________________________________________ BESS mailing list BESS@ietf.org https://www.ietf.org/mailman/listinfo/bess
