Jakob Homan commented on GIRAPH-45:

Another consideration is, in situations where a combiner has been defined and 
has a significant effect, that the destination worker is likely to have more 
messages than any other node, so by shipping them to it early, we'll save more 
space by running the combiner against all those messages.  ie, if four workers 
each have 10 messages and combine each of those sets to 1 message, we still 
have four messages.  If we then immediately ship them to the destination worker 
and combine, we would end up with 1 final message to be stored for the next 
> Improve the way to keep outgoing messages
> -----------------------------------------
>                 Key: GIRAPH-45
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-45
>             Project: Giraph
>          Issue Type: Improvement
>          Components: bsp
>            Reporter: Hyunsik Choi
>            Assignee: Hyunsik Choi
> As discussed in GIRAPH-12(http://goo.gl/CE32U), I think that there is a 
> potential problem to cause out of memory when the rate of message generation 
> is higher than the rate of message flush (or network bandwidth).
> To overcome this problem, we need more eager strategy for message flushing or 
> some approach to spill messages into disk.
> The below link is Dmitriy's suggestion.
> https://issues.apache.org/jira/browse/GIRAPH-12?focusedCommentId=13116253&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13116253

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: 
For more information on JIRA, see: http://www.atlassian.com/software/jira


Reply via email to