Hello Santosh,
  Thanks for your careful review. You bring up a valid point, that needs 
clarification. Even without multiple BFD sessions on the <FEC, LSP> if the 
egress were to inform about a BFD session going to DOWN (potentially using the 
diag code 0x4 - Neighbor Signalled Down), the YourDisc field needs to be 
non-zero, so the ingress can demux the packet to the correct session. With 
multiple sessions per FEC, LSP, this becomes absolutely needed, as you remark 
below. Could we consider adding the following statement:
The egress MUST use the discriminators exchanged when the session was brought 
to indicate a session DOWN back to the ingress. The egress SHOULD reset this to 
zero after transmitting detectMult number of packets.
Please suggest your thoughts on making the above text better.
Thanks
Prasad
Copying the list for their inputs as well.

-----Original Message-----
From: Santosh P K [mailto:[email protected]] 
Sent: Thursday, May 14, 2015 7:44 PM
To: [email protected]
Cc: Jeffrey Haas; Nobo Akiya
Subject: RE: Working Group Last Call for draft-ietf-bfd-rfc5884-clarifications

Hello Authors,
     Sorry for being so late in the game. I had few points which are not clear 
from draft. 


As per RFC 5880 section 6.8.1 it says that when BFD session goes down due to 
control timer expiry then remote discr MUST be set to 0. 

<SNIP>
"   bfd.RemoteDiscr

      The remote discriminator for this BFD session.  This is the
      discriminator chosen by the remote system, and is totally opaque
      to the local system.  This MUST be initialized to zero.  If a
      period of a Detection Time passes without the receipt of a valid,
      authenticated BFD packet from the remote system, this variable
      MUST be set to zero."

<SNIP>

As per RFC 5884 it says that all the demux MUST happen with remote 
Discriminator and also it talks about Discr not changing only when session is 
UP.

<SNIP>
Section 7:

   Note that once the BFD session for the MPLS LSP is UP, either end of
   the BFD session MUST NOT change the source IP address and the local
   discriminator values of the BFD Control packets it generates, unless
   it first brings down the session.
<SNIP>


As per  clarification draft " draft-ietf-bfd-rfc5884-clarifications" section 
2.4 iterates the same statement. 

<SNIP>
Section 2.4:

The discriminators of a BFD session established over an MPLS LSP
   cannot be changed when it is in UP state.
<SNIP>


Problem:

Now for example if we have multiple LSP's with the same FEC  then at egress it 
might be sending packet to same designation address. In this case if one of the 
BFD session goes down on egress then it would send BFD packet with down state 
with 0 discr. Seeing this ingress would try to demux with IFL and source 
address on the packet, this will lead to multiple sessions. I think this needs 
to be clarified as well.

I am not sure if this point is already covered or not. Do you think it is worth 
clarifying this point? I think it plays very important role in bringing down 
the BFD session in MPLS environment. Any thoughts? 


Thanks
Santosh P K 

> -----Original Message-----
> From: Rtg-bfd [mailto:[email protected]] On Behalf Of Jeffrey 
> Haas
> Sent: Tuesday, May 05, 2015 1:28 AM
> To: [email protected]
> Cc: [email protected]
> Subject: Working Group Last Call for 
> draft-ietf-bfd-rfc5884-clarifications
> 
> The BFD mailing list has been quiet since shortly after IETF 92.  We 
> have a number of documents that seem stable and ready to advance.
> 
> This email begins a two week Working Group Last Call for 
> draft-ietf-bfd- rfc5884-clarifications-01, Clarifications to RFC 5884.
> 
> This last call will end on May 15.
> 
> Simultaneously, the authors of this draft should state to the mailing 
> list whether there is any IPR associated with this draft or not.
> 
> This WGLC is also being cc'd to mpls@ietf to permit additional 
> commentary

Reply via email to