[
https://issues.apache.org/jira/browse/CASSANDRA-15393?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16967116#comment-16967116
]
Blake Eggleston commented on CASSANDRA-15393:
---------------------------------------------
Here's a rough initial pass at converting cells from buffers to arrays:
[https://github.com/bdeggleston/cassandra/tree/15387-arrays]
There are {{FIXME}}s everywhere and I still need to clean up some ideas that
didn't pan out and take another look at cell cloning, but +902/-402 is smaller
than I was expecting.
I've also started playing around with converting from buffer to a {{Value}}
type that can be backed by ByteBuffer, byte[], or native memory. It's nicer to
look at, but is also a much deeper rabbit hole. There is also code in a lot of
places that relies on ByteBuffers having mutable position that would need to
reworked and verified.
> Add byte array backed cells
> ---------------------------
>
> Key: CASSANDRA-15393
> URL: https://issues.apache.org/jira/browse/CASSANDRA-15393
> Project: Cassandra
> Issue Type: Sub-task
> Components: Local/Compaction
> Reporter: Blake Eggleston
> Assignee: Blake Eggleston
> Priority: Normal
> Fix For: 4.0
>
>
> We currently materialize all values as on heap byte buffers. Byte buffers
> have a fairly high overhead given how frequently they’re used, and on the
> compaction and local read path we don’t do anything that needs them. Use of
> byte buffer methods only happens on the coordinator. Using cells that are
> backed by byte arrays instead in these situations reduces compaction and read
> garbage up to 22% in many cases.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
