TSV-ART review of draft-pd-dispatch-msrp-websocket

[Allison Mankin <[e-mail]>]

Hey, folks,

I've reviewed this draft (draft-pd-dispatch-msrp-websocket-13) as part of
the TSV Area Review Team, paying special attention to transport-related
concerns. Please take these as any other (belated) IETF last call comments,
addressing them in conjunction with the present IESG review.

Summary: This draft specifies the WebSocket sub-protocol for MSRP, the
SIP-based messaging protocol.   It is very similar to RFC 7118, WebSocket
as a Transport for SIP, which is already a PS.  It does not appear to pose
any transport-related danger, and is broadly ready for publication as a PS.

Although the draft looks ready to go from a transport point of view, I have
a couple of small questions:

Section 5

Does the second sentence in the following mean "impossible for a WebSocket
MSRP client to communicate directly with other MSRP clients"?  The
paragraph is confusing.

   WebSocket clients cannot receive WebSocket connections initiated by
   other WebSocket clients or WebSocket servers.  This means that it is
   impossible for an MSRP client to communicate directly with other MSRP
   clients.  Therefore, all MSRP over WebSocket messages MUST be routed
   via an MSRP WebSocket Server.


Section 8

bob.example.com:8145 occurs in many of the path examples.  Although I
notice it also occurs in the MSRP Relay RFC (4976) along with many other
unassigned port numbers from the User Space, I wonder if it could be
replaced with a port from the Dynamic space (49152-65535) - in this draft,
it's the only unassigned port, and this could be one RFC that doesn't
confuse port users IRL.

Thanks and good luck,

Allison