Early Review of draft-ietf-ippm-ioam-integrity-yang-04
review-ietf-ippm-ioam-integrity-yang-04-yangdoctors-early-ma-2026-01-05-00

Hi, all,

This is my early review of draft-ietf-ippm-ioam-integrity-yang. Note I am not
an expert in IOAM, my review focuses primarily on the YANG data model.

The document is well-organized and easy to follow. The YANG module has passed
the pyang and yanglint validation, and the JSON examples provided are valid as
well. However, I am slightly concerned about the current model design which
seems to have duplicate definitions, e.g., the int-preallocated-tracing-profile
container defines its own node-action while the base ietf-ioam module already
provides one inside preallocated-tracing-profile, if a server supports both
feature preallocated-trace and int-preallocated-trace.

I wonder whether the WG have considered a simper modelling approach, such as:
OLD:
augment /ioam:ioam/ioam:profiles/ioam:profile:
   +--rw int-preallocated-tracing-profile! {int-preallocated-trace}?
       +--rw node-action?   ioam-node-action
       +--rw trace-types
       |  +--rw use-namespace?   ioam-namespace
       |  +--rw trace-type*      ioam-trace-type
       +--rw max-length?    uint32
       +--rw int-method?    iana-ioam-ipm:method-id
NEW:
 augment
 /ioam:ioam/ioam:profiles/ioam:profile/ioam:preallocated-tracing-profile:
    +--rw int-method?   iana-ioam-ipm:method-id {int-preallocated-trace}?

Or if the intent is to make a distinction, please consider adding constraints
to make sure containers like preallocated-tracing-profile and
int-preallocated-tracing-profile are mutually exclusive.

Sec.3.2 to sec.3.5, each subsection defines the int-method data node, and the
draft says it indicates which integrity protection method is used, but its type
definition uses the ID for the method and only includes method-0 without
specifying its semantic details, which may cause implementation ambiguities,
perhaps make this clearer and consider explicitly adding a reference where the
integrity protection method is formally defined.

Additionally, the int-method node is only effective when the node action is
action-encapsulate. Could you please clarify in the draft why it is not
required for other actions (e.g., action-transit)?

Best Regards,
Qiufang