[
https://issues.apache.org/jira/browse/AURORA-953?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14230475#comment-14230475
]
Kevin Sweeney commented on AURORA-953:
--------------------------------------
We could add a bool tail [default=false] field to FrameChunk to tag it. The
data fields would remain the same.
> Change the log format so that FrameHeader doesn't need to know the size and
> checksum of the serialized data ahead-of-time.
> --------------------------------------------------------------------------------------------------------------------------
>
> Key: AURORA-953
> URL: https://issues.apache.org/jira/browse/AURORA-953
> Project: Aurora
> Issue Type: Story
> Components: Scheduler
> Affects Versions: 0.6.0, 0.5.0
> Reporter: Kevin Sweeney
>
> If we write FrameHeader as a trailer we can get additional streaming
> optimizations that reduce the amount of heap the scheduler needs to operate.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
