RE: [ws-rx] NEW ISSUE: retransmission

[Doug Davis <dug@us.ibm.com>]

Well, since Acks can be received out
of order I don't think you want to say "most recent".
I think sticking with just talking about
"unacked" messages is safer.
thanks,
-Doug

"Gilbert Pilz" <Gilbert.Pilz@bea.com>

11/02/2006 07:35 PM

To	"Paul Fremantle" <paul@wso2.com>, <ws-rx@lists.oasis-open.org>
cc
Subject	RE: [ws-rx] NEW ISSUE: retransmission

May I suggest ammending your proposal in the following
way:

Change line 230 to read:

"While the Sequence is not closed or terminated, the following
invariants are REQUIRED for correctness:"

Then change the new bullet to be inserted after line 238 to read:

"The RMS MUST retransmit any messages that are missing from the most
recent Acknowledgement Message".

- gp

" . . and nice red uniforms."

> -----Original Message-----
> From: Paul Fremantle [mailto:paul@wso2.com] 
> Sent: Thursday, November 02, 2006 3:40 PM
> To: ws-rx@lists.oasis-open.org
> Subject: [ws-rx] NEW ISSUE: retransmission
> 
> We do not normatively state that any messages must be 
> retransmitted unless the server Nacks them.
> 
> Since the Protocol Invariants are there to explain how we 
> actually ensure reliable transmission, that is the 
> appropriate place to add this.
> 
> Proposal:
> 
> Add a new invariant:
> While the Sequence is not closed or terminated, the RMS must 
> retransmit any messages that are missing from the most recent 
> acknowledgement message.
> 
> 
> 
> -------- Original Message --------
> Subject:              
   [ws-rx] Potential new issue: retransmission
> Date:                
 Wed, 01 Nov 2006 12:22:25 +0000
> From:                
 Paul Fremantle <paul@wso2.com>
> Organisation:              
   WSO2
> To:                
 ws-rx@lists.oasis-open.org <ws-rx@lists.oasis-open.org>
> 
> 
> 
> While looking at PR016, I could only find three places where we 
> normatively define retransmission:
> 
> 1. Upon receipt of a NACK, you must retransmit that message
> 2. Upon MessageNumberRollover, you must continue to 
> retransmit messages
> 3. In the state tables, we have a state corresponding to this.
> 
> Given that NACK is optional, MessageNumberRollover highly unlikely,
> doesn't seem like we've defined this very well!
> 
> How about adding as a protocol invariant that the RMS must retransmit
> unacknowledged messages?
> 
> 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
> 
> 
>