I was looking for something stronger - such as the language from the CoAP draft
The following rules are added for matching a response to a request:
The DTLS session MUST be the same, and the epoch MUST be the same.
I could not find the language as I was searching for similar language dealing with matching. The language is adequate, but something that matches the other draft would still have been my preference.
Jim
-----Original Message-----
From: Carsten Bormann [mailto:***@tzi.org]
Sent: Thursday, May 18, 2017 9:35 AM
To: Jim Schaad <***@augustcellars.com>
Cc: draft-ietf-core-coap-tcp-***@ietf.org; ***@ietf.org
Subject: Re: [core] draft-ietf-core-coap-tcp-tls
Post by Jim SchaadThis document is missing any statements about making sure that the
same session (TLS, TCP, WebSockets,...) is used when matching requests
and replys. While this may seem to be obvious, the statement does
exist for DTLS and thus should be echoed here.
Hi Jim,
3.3 (for the TCP and TLS/TCP cases) says:
Responses MUST be returned over the same connection as
the originating request.
and goes on:
Concurrent requests are differentiated by
their Token, which is scoped locally to the connection.
4.3 (for the WebSockets cases) redundantly(*) says:
Responses MUST be returned over the same
connection as the originating request.
and also goes on:
Concurrent requests are
differentiated by their Token, which is scoped locally to the
connection.
Is there more we need to say?
Grüße, Carsten
(*) Para 1 of Section 4 says "this section only specifies the differences between the
transports.”. But I think it does not hurt to state this important point again.