Early Review of draft-ietf-opsawg-discardmodel-03
review-ietf-opsawg-discardmodel-03-yangdoctors-early-lhotka-2024-08-28-00

The reviewed Internet-Draft introduces YANG module
“ietf-packet-discard-reporting” that defines a hierarchical schema for
reporting received, transmitted and discarded network packets along with a
detailed classification of reasons for packet discard.

Whilst the module is relatively long, it is pretty trivial from the YANG point
of view in that it defines just the data hierarchy and no additional “business
rules” expressed e.g. via features or “when”/“must” statements.

It is not clear to me why the YANG module is touted as a representation of an
*information* model. Does it mean that an ensuing data model will contain more
details or be different in another way?

Perhaps related to the previous question is this: what is the reason for using
the data structure extension per RFC 8791? The I-D text indicates that the
schema could eventually be implemented in network devices for reporting packet
statistics, so “ietf-packet-discard-reporting” could IMO simply define regular
state data.

In many places, the module defines sibling containers “v4” and “v6” with
identical contents. This repetitive character is properly handled by YANG
groupings but, from the design point of view, I would prefer using a list with
address family as the key — using either “address-family” type from
“iana-routing-types” (RFC 8294) or an ad hoc type. This would not only make the
schema tree shorter but also make the module potentially applicable to other
address families. If there is a reason for using the hard-coded containers, it
should be stated in the I-D.

Nit: The module revisions in the “revision” statement and the file name in
“<CODE BEGINS>” differ.