You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Aljoscha Krettek (JIRA)" <ji...@apache.org> on 2016/02/08 13:14:40 UTC

[jira] [Resolved] (FLINK-3336) Add Semi-Rebalance Data Shipping for DataStream

     [ https://issues.apache.org/jira/browse/FLINK-3336?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Aljoscha Krettek resolved FLINK-3336.
-------------------------------------
    Resolution: Fixed

Added in https://github.com/apache/flink/commit/9f6a8b6d075551d1b4090e191df8f56b5ef9d7e9

> Add Semi-Rebalance Data Shipping for DataStream
> -----------------------------------------------
>
>                 Key: FLINK-3336
>                 URL: https://issues.apache.org/jira/browse/FLINK-3336
>             Project: Flink
>          Issue Type: Improvement
>          Components: Streaming
>            Reporter: Aljoscha Krettek
>            Assignee: Aljoscha Krettek
>             Fix For: 1.0.0
>
>
> This feature has recently been requested on the ML: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/Distribution-of-sinks-among-the-nodes-td4640.html
> The new data shipping pattern would allow to rebalance data only to a subset of downstream operations.
> The subset of downstream operations to which the upstream operation would send
> elements depends on the degree of parallelism of both the upstream and downstream operation.
> For example, if the upstream operation has parallelism 2 and the downstream operation
> has parallelism 4, then one upstream operation would distribute elements to two
> downstream operations while the other upstream operation would distribute to the other
> two downstream operations. If, on the other hand, the downstream operation had parallelism
> 2 while the upstream operation has parallelism 4 then two upstream operations would
> distribute to one downstream operation while the other two upstream operations would
> distribute to the other downstream operations.
> In cases where the different parallelisms are not multiples of each other one or several
> downstream operations would have a differing number of inputs from upstream operations.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)