Note that I'm behind in this thread, but wanted to make comment on the 
specifically raised point:


> On Oct 29, 2025, at 11:33 AM, Rao P, Gopinatha 
> <[email protected]> wrote:
> 
> Hi Venkat,
>  
> From RFC 5880:
>  
>       “If the Your Discriminator field is zero, the session MUST be
>       selected based on some combination of other fields, possibly
>       including source addressing information, the My Discriminator
>       field, and the interface over which the packet was received.  The
>       exact method of selection is application specific and is thus
>       outside the scope of this specification.  If a matching session is
>       not found, a new session MAY be created, or the packet MAY be
>       discarded.  This choice is outside the scope of this
>       specification.”
>  
> When the remote PE sends a DOWN packet to begin with, your_discriminator can 
> be 0 and the incoming BFD packet can be demuxed based on the VNI(and other 
> parameters) on which the packet is coming in on ? RFC 5880 doesn’t mandate 
> the incoming DOWN packet should come in with your_discriminator set to 
> non-zero value. It clearly says what needs to be done when your_discriminator 
> field to set to 0.

The followup on this is that if you're getting such a 0 discriminator, the idea 
is that you have other mechanisms to determine how to associate these packets 
with a BFD session.  The typical simple example is for single-hop BFD you can 
see what interface and IP address things are coming in as.

RFC 5882 for multihop both acknowledges use cases like demultiplexing based on 
IP, but also that sometimes you need to learn this out of band.

The question for the draft authors and the BESS WG is whether there's 
alternative means that such a EVPN BFD session could use to associate incoming 
packets to its session?  If no, then a "MUST" for discovered (by BGP) or 
provisioned is reasonable.  Clarifying that this is required since there's no 
alternative discovery mechanism available might address the raised point.

-- Jeff

_______________________________________________
BESS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to