[
https://issues.apache.org/jira/browse/RATIS-1312?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17294620#comment-17294620
]
runzhiwang edited comment on RATIS-1312 at 3/3/21, 3:58 PM:
------------------------------------------------------------
> 1. decodeDataStreamRequestByteBuf costs 6.42% caused by, netty splits 4MB to
> many small packets, each about 60KB, ...
[~szetszwo] I think 60KB is the TCP packet size. But IMOP, I think netty should
be responsible for collecting all the packets belongs to buffer size such 4MB,
because we send 4MB each time, when netty receive all packets of 4MB, it call
back ratis, and we should not check whether we have receive all packets by
checking h.getDataLength() + headerBufLen <= buf.readableBytes() when we
receive each packet, it's not ratis's responsibility.
Simply to say, client sends 4MB to netty, when netty callback ratis, netty
should pass 4MB to ratis, not 60KB packet one by one.
was (Author: yjxxtd):
> 1. decodeDataStreamRequestByteBuf costs 6.42% caused by, netty splits 4MB to
> many small packets, each about 60KB, ...
[~szetszwo] I think 60KB is the TCP packet size. But IMOP, I think netty should
be responsible for collecting all the packets belongs to buffer size such 4MB,
because we send 4MB each time, when netty receive all packets of 4MB, it call
back ratis, and we should not check whether we have receive all packets by
checking h.getDataLength() + headerBufLen <= buf.readableBytes() when we
receive each packet, it's not ratis's responsibility.
> Compare the performance between HDFS and DataStreamApi
> ------------------------------------------------------
>
> Key: RATIS-1312
> URL: https://issues.apache.org/jira/browse/RATIS-1312
> Project: Ratis
> Issue Type: Sub-task
> Reporter: runzhiwang
> Priority: Major
> Attachments: hdfs.svg, streaming.svg
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
