[
https://issues.apache.org/jira/browse/CASSANDRA-4548?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis updated CASSANDRA-4548:
--------------------------------------
Attachment: 4548.txt
attached. (the boolean isn't used either, we don't bother sending responses on
failure, letting them time out.)
> Mutation response(WriteResponse.java) could be smaller and not contain
> keyspace and key
> ---------------------------------------------------------------------------------------
>
> Key: CASSANDRA-4548
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4548
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: sankalp kohli
> Assignee: Jonathan Ellis
> Priority: Minor
> Labels: network
> Fix For: 1.2.0 beta 1
>
> Attachments: 4548.txt
>
>
> In the mutation response, WriteResponse.java object is send back to the
> co-ordinator. This object has keyspace and key in it which is not required.
> It is not being used at the co-ordiantor.
> This wastes IO specially in case of WAN links between DC. Also since response
> from each node in multi-DC deployments goes back to the co-ordinator in
> another DC makes it even worse.
> It also becomes worse if the the keyspace and key are of large size and the
> data is small. In that case, a node which is not the co-ordinator and purely
> receiving mutations, the outbound n/w bandwidth could be half of incoming
> bandwidth.
--
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
