← Prev in month ← Prev in thread

Review: AMQP Addressing Version 1.0 (WD10)

From
Rob Godfrey <>
Date
2019-10-09T12:32:34+00:00
ID
CAA2jDtXM+vxjJiJYs3Z=0pAJLyv0FupvcLvE=
Thread
Review: AMQP Addressing Version 1.0 (WD10)
Section 3.1 Protocol Schemes

I think "AMQP WebSocket Binding [AMQPWS] endpoints MUST either be described with the standard ws (non-secure) or wss (secure, TLS) WebSocket schemes." would read more clearly as "AMQP WebSocket Binding [AMQPWS]
endpoints MUST be described with the standard ws (non-secure) or wss (secure, TLS) WebSocket schemes." (i.e. remove 'either')

Section 3.2.2 Message 'to' field
Section 3.2.3 Message 'reply-to' and the Request Reply pattern

In these sections we refer to notions of "receiver"s, "router"s and "server"s. These terms are introduced without definition.

In section 3.2.2 in particular we have:

Any receiver MUST ignore the network
endpoint and not use it as a dispatch criterion, as access via different
on-ramps to the same AMQP address is equivalent.

A router MAY ignore the 'to' field and
forward a message within its own network, or MAY connect to the 'to' field network
address. How routers decide that is out of scope of this specification; they
MAY use custom annotations, properties of the link or connection that received
the message, or other mechanisms.

which is specifying different behaviour of receivers and routers.

I think we need another working draft to resolve this before we can progress to CSPRD.

-- Rob
-- 

_____________________________________________________________________________Red Hat GmbH,Âwww.de.redhat.com,Registered seat: Grasbrunn, Commercial register: Amtsgericht Muenchen, HRB 153243,Managing Directors: ,Charles Cachera, Michael Cunningham, Michael O'Neill, Eric Shander
← Prev in month ← Prev in thread