+1
Another useful thing this seems to prevent is sending an
AckRequested with a piggybacked ack for an offered sequence.
I don’t see the point in restricting what messages acks can be
piggybacked on beyond requiring those messages to be addressed at AcksTo. That
should be up to some binding that profiles usage of the protocol.
--Stefan
From: Doug Davis
[mailto:dug@us.ibm.com]
Sent: Thursday, November 30, 2006 12:40 PM
To: Paul Fremantle
Cc: ws-rx@lists.oasis-open.org
Subject: Re: [ws-rx] Proposal for i021
Still thinking
about this one but my initial thought is that the AcksTo EPR is specifically
called the "AcksTo" EPR for a reason - there must be an RM processor
there to handle Acks. So, I'm not really sure any restrictions are needed
- this goes to your first paragraph.
To your
specific text - your first bullet would prevent us from optimizing things and
putting Acks for lots of sequences onto the same message.
You're 2nd
bullet (I think) is related to this in that it seems to acknowledge that we may
want to put multiple acks for multiple sequences into the message and we're
just concerned with knowing that some RM processor is at the EPR - well, then
we're back to my first point - if there wasn't an RM processor there then the
AcksTo EPR is bad.
thanks
-Doug
__________________________________________________
STSM | Web Services Architect | IBM Software Group
(919) 254-6905 | IBM T/L 444-6906 | dug@us.ibm.com
Paul
Fremantle <paul@wso2.com>
11/30/2006
02:34 PM
|
To
|
"ws-rx@lists.oasis-open.org"
<ws-rx@lists.oasis-open.org>
|
cc
|
|
Subject
|
[ws-rx]
Proposal for i021
|
|
The high-level view is that it would be nice
to restrict the
piggybacking of acks to situations where we are sure that there is an RM
agent at the other end.
Replace the beginning of section 3.9 with the following:
The RM Destination informs the RM Source of successful message receipt
using a
SequenceAcknowledgement header block. The RM Destination MAY Transmit the
SequenceAcknowledgement header block independently or it MAY include the
SequenceAcknowledgement header block on existing messages targeted to
the AcksTo EPR.
When the SequenceAcknowledgement header block is included on existing
messages, this is known as piggybacking. Piggybacking MAY occur in two
cases:
* The first case is where the SequenceAcknowledgement header block
is piggybacked onto a reply to a Sequence Traffic Message. In
this
case the SequenceAcknowledgement must apply to the same
Sequence
as the SequenceTrafficMessage. The definition of reply used
is
that defined by the WS-Addressing relationship URI
"http://www.w3.org/2005/08/addressing/reply".
* The second case is where the existing message is a Sequence
Traffic Message. In this case the Sequence of the Sequence
Traffic
Message does not need to be the same as the Sequence of the
SequenceAcknowledgement header block.
Piggybacking MUST not occur unless one or both of these cases apply.
Paul
--
Paul Fremantle
VP/Technology and Partnerships, WSO2
OASIS WS-RX TC Co-chair
http://bloglines.com/blog/paulfremantle
paul@wso2.com
(646) 290 8050
"Oxygenating the Web Service Platform", www.wso2.com
--
Paul Fremantle
VP/Technology and Partnerships, WSO2
OASIS WS-RX TC Co-chair
http://bloglines.com/blog/paulfremantle
paul@wso2.com
(646) 290 8050
"Oxygenating the Web Service Platform", www.wso2.com