[
https://issues.apache.org/jira/browse/KAFKA-8452?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16852540#comment-16852540
]
ASF GitHub Bot commented on KAFKA-8452:
---------------------------------------
vvcephei commented on pull request #6848: KAFKA-8452: Compressed BufferValue
URL: https://github.com/apache/kafka/pull/6848
De-duplicate the common case in which the prior value is the same as the old
value.
### Committer Checklist (excluded from commit message)
- [ ] Verify design and implementation
- [ ] Verify test coverage and CI build status
- [ ] Verify documentation (including upgrade notes)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Possible Suppress buffer optimization: de-duplicate prior value
> ---------------------------------------------------------------
>
> Key: KAFKA-8452
> URL: https://issues.apache.org/jira/browse/KAFKA-8452
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: John Roesler
> Priority: Major
>
> As of KAFKA-8199, the suppression buffers have to track the "prior value" in
> addition to the "old" and "new" values for each record, to support
> transparent downstream views.
> In many cases, the prior value is actually the same as the old value, and we
> could avoid storing it separately. The challenge is that the old and new
> values are already serialized into a common array (as a Change via the
> FullChangeSerde), so the "prior" value would actually be a slice on the
> underlying array. But, of course, Java does not have array slices.
> To get around this, we either need to switch to ByteBuffers (which support
> slices) or break apart the serialized Change into just serialized old and new
> values.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
