[
https://issues.apache.org/jira/browse/KAFKA-5285?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16295649#comment-16295649
]
Guozhang Wang commented on KAFKA-5285:
--------------------------------------
I'll let [~damianguy] to chime in more on this issue. Though I think [~xvrl]'s
reported one and [~davispw]'s reported one are correlated but not exactly the
same.
My understanding is that, for session windows where the session's ending
timestamps cannot be defined until the time the window is closed / merged,
because of the way we lay out the data in the underlying key value store, we
have to be conservative to determine the range that we need to search for
applicable sessions. I do not have a concrete idea on how to improve this
observed issues, but if you guys have some idea on how to improve, let's
discuss them here.
> optimize upper / lower byte range for key range scan on windowed stores
> -----------------------------------------------------------------------
>
> Key: KAFKA-5285
> URL: https://issues.apache.org/jira/browse/KAFKA-5285
> Project: Kafka
> Issue Type: Improvement
> Components: streams
> Reporter: Xavier Léauté
> Assignee: Xavier Léauté
> Labels: performance
>
> The current implementation of {{WindowKeySchema}} / {{SessionKeySchema}}
> {{upperRange}} and {{lowerRange}} does not make any assumptions with respect
> to the other key bound (e.g. the upper byte bound does not depends on lower
> key bound).
> It should be possible to optimize the byte range somewhat further using the
> information provided by the lower bound.
> More specifically, by incorporating that information, we should be able to
> eliminate the corresponding {{upperRangeFixedSize}} and
> {{lowerRangeFixedSize}}, since the result should be the same if we implement
> that optimization.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
