YANG Data Model for Bidirectional Forwarding Detection (BFD)

Note: This ballot was opened for revision 15 and is now closed.

Martin Vigoureux Yes

(Ignas Bagdonas) No Objection

Deborah Brungard No Objection

(Ben Campbell) No Objection

Alissa Cooper (was Discuss) No Objection

Comment (2018-07-19 for -16)
No email
send info
Looking forward to the discussion of my original DISCUSS in NETMOD. Here it was:

I will apologize in advance because this document may be sort of a casualty of this DISCUSS. I should have raised my point below at least two years ago if not four years ago when the first iana-* YANG module was registered, but the thought did not occur to me until now. 

It gives me some pause to see the name "iana" embedded in the file name, module name, namespace, and prefix of the module being defined in Sec. 2.12. I realize there is precedent here, but I question whether tying these kinds of modules specifically to IANA as the protocol parameter registry operator by name puts them on the most stable deployment footing under all possible circumstances. I am personally pleased as punch with the service we get from IANA, but that doesn't mean "IANA" will always be the name of the registry operator. The more modules that get created with this embedding, the more of them that may need to change in the unlikely event that the name of the registry operator changes. Lots of RFCs would need to change too, but embedding the name extends the potential problem to the modules themselves.

It wasn't clear to me whether there is some ops-area-wide convention around the embedding of "iana" in the names of modules to be maintained by IANA. I don't see this specifically referenced in RFC 7950 or RFC 6020. So I'd like to discuss whether a different naming convention could be established and used in this document and others going forward.



Some further questions on Sec. 2.12:

Looking back at the other RFCs that have defined YANG modules to be maintained by IANA (7224, 7317, 8294, 8348), they use two different postal addresses for ICANN. Why? Furthermore, is "ICANN" really the right contact name, or should it be PTI?

Benjamin Kaduk (was Discuss) No Objection

Comment (2018-08-02)
No email
send info
Thank you for your patience and continued discussion as we worked to resolve my DISCUSS point!

(Suresh Krishnan) No Objection

Warren Kumari (was Discuss) No Objection

Comment (2018-07-04 for -16)
No email
send info
Thank you.

I also had a few minor nits:
Section 1:
"The YANG modules in this document conform to the Network Management Datastore Architecture (NMDA) Network Management Datastore Architecture [RFC8342]. " 
The Department of Redundancy Department called and wants some of their words back please :-)

Section 2:
"Since BFD is used for liveliness detection of various forwarding
   paths, there is no uniform key to identify a BFD session.  So the BFD
   data model is split in multiple YANG modules where each module
   corresponds to one type of forwarding path."
I think this would be more readable as:
"... to identify a BFD session, and so the BFD..."  (hey, I said it was a nit)

(Mirja K├╝hlewind) No Objection

(Terry Manderson) No Objection

(Alexey Melnikov) No Objection

(Eric Rescorla) No Objection

Comment (2018-07-04 for -16)
No email
send info
Rich version of this review at:

S 2.1.4.
>              Minimum TTL of incoming BFD control packets.
>   2.1.4.  MPLS Traffic Engineering Tunnels
>      For MPLS-TE tunnels, BFD is configured under the MPLS-TE tunnel since
>      the desired failure detection parameters is a property of the MPLS-TE

"parameters are"

S 2.8.
>   2.8.  BFD over LAG hierarchy
>      A "lag" node is added under the "bfd" node in control-plane-protocol.
>      The configuration and operational state data for each BFD LAG session
>      is under this "lag" node.

There seems to be a lot of replication (e.g., number of sessions). Is
it possible to somehow refactor this so that's common?

Alvaro Retana No Objection

(Adam Roach) No Objection

Comment (2018-07-03 for -16)
No email
send info
I have found one minor editorial issue on page 43:

>    notification singlehop-notification {
>      description
>        "Notification for BFD single-hop session state change. An " +
>        "implementation may rate-limit notifications, e.g. when a" +
>        "session is continuously changing state.";

Nit: add a space between "a" and the closing quotation mark.

(Note: this occurs on Pages 46, 50, 54, and 57 as well)