A YANG Data Model for the Multicast Source Discovery Protocol (MSDP)
draft-ietf-pim-msdp-yang-18

Random questions:

* If the thing being modelled is experimental (3618) should this also be
  experimental (as opposed to standards track)?

I'm assuming the sponsoring AD approved the presence of six authors on the document, exceeding our normal limit of five.

I'm a YANG newbie so for the moment I'm going to defer to those more expert at evaluating the syntax and semantics of the model described in Sections 3 and 4 while I go learn more about YANG in general.

Thus, the only feedback I have is editorial stuff in Section 5:
* "Modifying the configuration can cause MSDP default peers to be deleted or the connection to be rebuilt, and the SA's unexpected filtering." -- does that last clause mean "unexpected filtering by the SA"?
* "The key misoperation will broke the existed MSDP connection ... " -- s/broke the existed/break the existing/, right?
* "These are the subtrees and data nodes and their sensitivity/vulnerability:" followed by a list of one subtree/node.  Maybe just turn that into prose and name the singular subtree?
* "... may lead to forged connection attack, ..." -- s/to/to a/
* "... the ACL nodes uncorrected modification may lead to the filter errors." -- "nodes" should be "node's" or "nodes'"
* "The unauthorized reading function ..." -- perhaps "Unauthorized reads"?
* "... allow the unexpected peer connection rebuilding." -- maybe "allow the unexpected rebuilding of connection peers"?
* "Authentication configuration is supported ..." -- feels like this should start a new paragraph

I support Ben Kaduk’s DISCUSS.  The binding between the authentication container and MSDP isn’t clear.

I only did a quick scan of this document.

I am not sure whether there are any I18N considerations for this leaf:
          case password {
            leaf key {
              type string;
              description
                "This leaf describes the authentication key.";
              reference
                "RFC 8177: YANG Data Model for Key Chains.";
            }

I also support Ben’s DISCUSS.

Section 1.2 gives the BCP 14 boilerplate and references, but there are no BCP 14 key words in the document.  I suggest removing the boilerplate and the references.

Thanks for the updates; they seem to address the core issues raised in my DISCUSS points.
That said, I do still have a couple additional requested changes.

With respect to authentication, I don't think we say much about how having an
authentication-type configured implies that the corresponding authentication type
is to be used at runtime (i.e., TCP-MD5 given the current state of things), and contrariwise
that the absence of a configured authentication type implies that unauthenticated TCP is used.

Also, for the leaf descriptions:

"up-time": "Indicates the duration of time since this SA entry was created in the cache. [...]"
"expire": "Indicates the duration of time until this SA entry in the cache will time out. [...]"

Thanks for the updates.