[
https://issues.apache.org/jira/browse/CASSANDRA-9894?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14679960#comment-14679960
]
Benedict commented on CASSANDRA-9894:
-------------------------------------
I've rolled back the last commit, since it was less efficient, and decided it
is not worth chasing a cleaner version. It's possible to make serialization
better for huge sets, but not sufficiently to be worth the investment (given we
expect just about never to encounter this serialization path in the wild for
such huge sets, and if we do this probably won't be the botleneck).
While looking over the code, I also realised there is a bug in
{{selectOrderIterator}}, and realised then I had misread your
{{selectOrderIterator}} comment as referring to {{inOrderInclusionTester}}. So,
well, lesson learned I guess. Or hope. I've fixed this, and added some comments
associated to its use and added some suitable testing.
> Serialize the header only once per message
> ------------------------------------------
>
> Key: CASSANDRA-9894
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9894
> Project: Cassandra
> Issue Type: Sub-task
> Components: Core
> Reporter: Sylvain Lebresne
> Assignee: Benedict
> Fix For: 3.0 beta 1
>
>
> One last improvement I'd like to do on the serialization side is that we
> currently serialize the {{SerializationHeader}} for each partition. That
> header contains the serialized columns in particular and for range queries,
> serializing that for every partition is wasted (note that it's only a problem
> for the messaging protocol as for sstable we only write the header once per
> sstable).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
