<Cross-posting to BFD>
Hi Andrew, authors,
I had concern that this document may violate S-BFD spec but upon checking 
RFC7880 section 7.2.1, it seems that there is no issue. BTW I think RFC7881 
should also be in the reference list.
I had my colleague Claude do a quick check and Claude believes there's an 
issue. I am too tired to verify this right now, but I am sending it just in 
case I forget.
Regards,Reshad.

Section 7.3.2 recommends that each egress SID list of each Junction Segment run 
"its own S-BFD session to the next downstream node." In the Appendix A example, 
Junction Segment C has SID List 2 [Adj-SID-CF, BSID-F]. The implied S-BFD 
session would be initiated from C and is supposed to verify reachability to F.

But that's not what happens. When an S-BFD packet (per RFC 7881 / RFC 9612) is 
forwarded along [Adj-SID-CF, BSID-F]:
   
   - C pushes the full label stack and forwards to F via the adjacency SID.
   - F receives the packet with BSID-F on top. F processes BSID-F as a Binding 
SID — it swaps BSID-F for its own Junction Segment SID list [Adj-SID-FH] and 
forwards to H.
   - The S-BFD packet terminates at H (the DAG egress), not at F.
_______________________________________________
spring mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to