[
https://issues.apache.org/jira/browse/CASSANDRA-8692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14299086#comment-14299086
]
jonathan lacefield commented on CASSANDRA-8692:
-----------------------------------------------
While performance testing with a user, we noticed good results after applying
this patch.
Our performance testing was a simple 50/50 read/write test using 2.1 stress
and the following artifacts:
Simple.yaml - stress test yaml file:
https://drive.google.com/file/d/0B4Imdpu2YrEbblVCVENvVTJIYjg/view?usp=sharing
Stress results -
https://docs.google.com/spreadsheets/d/13UFCqh-W2wK_kL97Mh_rPGw-RLv7dKvWDLKQ_SXq5Mo/edit?usp=sharing
The patch was applied after "run 25" in the stress results.
We were testing for SLA adherence to see how much throughput could be
achieved while staying within an SLA target.
After applying this patch, throughput increased from ~79k ops to ~94k ops
while maintaining the SLA target.
The cluster used for this exercise was a 6 node cluster hosted on ec2 on
c3.8xl vms.
> Coalesce intra-cluster network messages
> ---------------------------------------
>
> Key: CASSANDRA-8692
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8692
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Ariel Weisberg
> Assignee: Ariel Weisberg
> Attachments: batching-benchmark.png
>
>
> While researching CASSANDRA-8457 we found that it is effective and can be
> done without introducing additional latency at low concurrency/throughput.
> The patch from that was used and found to be useful in a real life scenario
> so I propose we implement this in 2.1 in addition to 3.0.
> The change set is a single file and is small enough to be reviewable.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
