[netconf] Re: WGLC for distributed-notif-10

[Qin Wu <[e-mail]>]

Hi, Thomas and all:
I  take some time to review the v-(10), in my view, it is well written. Below are a few comments to this draft:

1.       Section 2 Terminologies
The definition of most of terms give people confusion that they are not complete or lack subject. Here are a few suggested changes:
Old text:
“
   Global Subscription: is the Subscription requested by the subscriber.
   It may be decomposed into multiple Component Subscriptions.

   Component Subscription: is the Subscription that defines a data
   source which is managed and controlled by a single Publisher.

   Global Capability: is the overall subscription capability that the
   group of Publishers can expose to the Subscriber.

   Component Capability: is the subscription capability that each
   Publisher can expose to the Subscriber.

   Master: is the Publisher that interacts with the Subscriber to deal
   with the Global Subscription.  It decomposes the Global Subscription
   to multiple Component Subscriptions and interacts with the Agents.

   Agent: is the Publisher that interacts with the Master to deal with
   the Component Subscription and pushing the data to the Receiver.

   Node: is the Publisher that obtains and pushes the data to the
   Receiver.

”
New Text:
“
   Global Subscription: The Subscription requested by the subscriber.
   It may be decomposed into multiple Component Subscriptions.

   Component Subscription: The Subscription that defines a data
   source which is managed and controlled by a single Publisher.

   Global Capability: The overall subscription capability that the
   group of Publishers can expose to the Subscriber.

   Component Capability: The subscription capability that each
   Publisher can expose to the Subscriber.

   Master: The Publisher that interacts with the Subscriber to deal
   with the Global Subscription.  It decomposes the Global Subscription
   to multiple Component Subscriptions and interacts with the Agents.

   Agent: The Publisher that interacts with the Master to deal with
   the Component Subscription and pushing the data to the Receiver.

   Node: The Publisher that obtains and pushes the data to the
   Receiver.

”

2.       Section 2 Terminology
In addition, I feel Terms like “Agent” and “Master” are too generic, suggest to
Replace them with “Publisher Agent” ,“Publisher Master”, ”Publisher Node”.

3.       Section 3 said:
“

Lost and corrupt YANG notification messages need to be recognized at

   the receiver to ensure data integrity even when multiple publisher

   processes publishing from the same transport session.
”
What to if lost or corrupt YANG notification is recognized at the receiver?
Is there a need to allow receive to inform publishers about failure event


4.       Section 4 said:

“

The Agents announce the status of their Component Subscriptions to

      the Master.  The status of the overall subscription is maintained

      by the Master.



”

Later on section 7 said:

“

   In addition to sending event records to Receivers, the Master MUST

   also send subscription state change notifications [RFC8639] when

   events related to subscription management have occurred.  All the

   subscription state change notifications MUST be delivered by the

   Master.

”
Is there any connection between component subscription status in section 4 and subscription state change in section 7.


5.       Section 7 said:

“   When the subscription decomposition result changed, the

   "subscription-modified" notification MUST be sent to indicate the new

   list of Publishers.

”

Should we make clear here a list of publishers are a list of Publisher Agents?NO?

6.       Section 4 said:

“   *  Contracts are created between the Master and each Agent on the

      Component Capability, and the format for streaming data structure.
”
I am wondering whether communication between master and agent is stateless?
At least I can see mapping relationship between master and multiple agents need to be maintained at the master,

Hope these comments are helpful.

-Qin
发件人: Kent Watsen [mailto:[e-mail]]
发送时间: 2025年2月8日 2:48
收件人: [e-mail]
主题: [netconf] WGLC for distributed-notif-10

This email begins a two-week WGLC on:

       Subscription to Distributed Notifications
       https://datatracker.ietf.org/doc/draft-ietf-netconf-distributed-notif/10/

Please take time to review this draft and post comments by February 21.  Favorable comments are especially welcomed.  Authors do not need to reply.


FWIW, none of the authors or contributors declared IPR:

       https://mailarchive.ietf.org/arch/msg/netconf/WFWTWNu8lww4xqTNsid3X6kA7zw/

Kent  // co-chair