[ 
https://issues.apache.org/jira/browse/DISPATCH-781?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17100819#comment-17100819
 ] 

ASF GitHub Bot commented on DISPATCH-781:
-----------------------------------------

ted-ross commented on pull request #731:
URL: https://github.com/apache/qpid-dispatch/pull/731#issuecomment-624661094


   > One suggestion: The logic that adds/removes the link_ref to the 
qdr_address_t (QDR_LINK_CLASS_ADDRESS) by checking the initial_credit_received 
value is repeated across a couple of files here.
   > 
   > I'd suggest moving that logic into its own two functions and calling them 
instead.
   
   Thanks.  I'll look into making this change.


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


> Tie end-to-end flow control of message-routed deliveries to outgoing capacity
> -----------------------------------------------------------------------------
>
>                 Key: DISPATCH-781
>                 URL: https://issues.apache.org/jira/browse/DISPATCH-781
>             Project: Qpid Dispatch
>          Issue Type: Improvement
>          Components: Router Node, Routing Engine
>            Reporter: Ted Ross
>            Assignee: Ted Ross
>            Priority: Major
>             Fix For: 1.13.0
>
>
> This feature reverses the management of incoming credit for message-routed 
> deliveries.
> The current mechanism is based on the the capacity of the incoming 
> (sender-client) links.  Each attached sender is given credit equal to the 
> configured link capacity regardless of the capacity of outgoing 
> (receiver-client) links for the same address.  This means that even under 
> congestion, a newly attached sender will get full capacity credit to send.
> The proposed mechanism is based not on incoming capacity but on outgoing 
> capacity.  In this case, the senders are provided a share of the total 
> outgoing capacity for an address.  As the number of incoming links (senders) 
> for an address changes and the total capacity for outgoing links changes, the 
> credit window of the senders shall be adjusted.
> The proposal is a heuristic approach.  It does not revoke credit from senders 
> (but will limit the number of credits that are replenished).  It will also 
> not deny credit if there are more incoming links than outgoing capacity.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org

Reply via email to