[jboss-jira] [JBoss JIRA] Commented: (JGRP-1091) NAKACK: further optimizations

[Bela Ban (JIRA)]

    [ https://jira.jboss.org/jira/browse/JGRP-1091?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12492831#action_12492831 ] 

Bela Ban commented on JGRP-1091:
--------------------------------

Here's another one: loop back multicast messages already in NAKACK (not only at the transport level).

In down(), if a message is multicast, add it to the xmit-table (no change to current behavior), but then try to remove as many messages as possible and send them up (in a separate thread).

This avoids concurrency between sending of messages by P and reception of P's messages on P's NakReceiverWindow.

> NAKACK: further optimizations
> -----------------------------
>
>                 Key: JGRP-1091
>                 URL: https://jira.jboss.org/jira/browse/JGRP-1091
>             Project: JGroups
>          Issue Type: Feature Request
>            Reporter: Bela Ban
>            Assignee: Bela Ban
>             Fix For: 2.8
>
>
> Sending of a message in down():
> - When sending a message in down(), we always grab the NakReceiverWindow for local_addr
> - A ref to the window can be cached, so we can eliminate the Map.get(local_addr) altogether

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira