Early Review of draft-ietf-opsawg-discardmodel-10
review-ietf-opsawg-discardmodel-10-yangdoctors-early-lhotka-2026-01-08-00

Request Review of draft-ietf-opsawg-discardmodel-10
Requested revision 10 (document currently at 12)
Type Early Review
Team YANG Doctors (yangdoctors)
Deadline 2026-01-12
Requested 2025-12-21
Requested by Joe Clarke
Authors John Evans , Oleksandr Pylypenko , Jeffrey Haas , Aviran Kadosh , Mohamed Boucadair
I-D last updated 2026-03-14 (Latest revision 2026-03-14)
Completed reviews Yangdoctors Early review of -03 by Ladislav Lhotka (diff)
Opsdir Early review of -09 by Sergio Belotti (diff)
Intdir Early review of -09 by Satoru Matsushima (diff)
Yangdoctors Early review of -10 by Ladislav Lhotka (diff)
Comments 
I'd like to request an updated YANG Doc review for this draft as it has entered WG LC, and the YANG has changed quite a bit to be both an IM and another DM module.  Moreover, there are NACM warning flagged by yanglint, and the authors could benefit from some renewed guidance from the YANG experts.  Thanks!
Assignment Reviewer Ladislav Lhotka
State Completed
Request Early review on draft-ietf-opsawg-discardmodel by YANG Doctors Assigned
Posted at https://mailarchive.ietf.org/arch/msg/yang-doctors/9fgDeAOGgAVJZAF4rbTwM4dKtVg
Reviewed revision 10 (document currently at 12)
Result Ready w/issues
Completed 2026-01-08
review-ietf-opsawg-discardmodel-10-yangdoctors-early-lhotka-2026-01-08-00
I've done an early review of version -03 of this draft, see
https://datatracker.ietf.org/doc/review-ietf-opsawg-discardmodel-03-yangdoctors-early-lhotka-2024-08-28/.
The authors addressed most of my comments and suggestions in the present
version.

**** General comments

The document now contains two YANG modules:

- ietf-packet-discard-reporting-sx contains two dozen groupings with components
of the schema for detailed packet loss statistics as well as several YANG
features and identities. This module then also defines a YANG structure [RFC
8791] that defines a complete hierarchical schema. This module – or rather the
structure – represents the information model (IM).

- ietf-packet-discard-reporting (the data model, DM) then augments three
existing modules (ietf-interfaces, ietf-logical-network-element and
ietf-routing) with relevant packet loss statistics from the IM.

This separation is useful and can indeed aid implementing the same data
structures in other protocols such as IPFIX. I would recommend to put the
"packet-discard-reporting" structure to a separate module. This way,
implementations that want to use the ietf-packet-discard-reporting module (and
thus have to import ietf-packet-discard-reporting-sx) needn't parse the
structure.

**** Specific comments

- The ietf-packet-discard-reporting module augments
"/rt:routing/rt:control-plane-protocols/rt:control-plane-protocol" with the
contents of the "plr-sx:control-plane" grouping. The augment target is a
"config true" data node and the grouping doesn't contain "config false", so the
augmented statistics are also "config true" (rw), which is probably wrong.

- What is the purpose of defining "alias" groupings that only use another
grouping and nothing else? This applies to the following pairs of groupings:
l2-traffic/basic-frames-bytes and l3-traffic/ip.

- I assume "all" identity means "either ipv4 or ipv6 address family". If it is
so, I'd suggest to use the following hierarchy of identities:

  address-family
    ip
      ipv4
      ipv6

  The intermediate "ip" identity would replace "all". This is more flexible
  because "ipv4" and "ipv6" are also derived from "ip" (unlike "all", and it
  also allows for adding other address families.