[
https://issues.apache.org/jira/browse/RATIS-1312?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17334572#comment-17334572
]
runzhiwang commented on RATIS-1312:
-----------------------------------
[~szetszwo] I think I find the reason why ratis streaming is slower than HDFS.
1. when client -> dn1 -> dn2 -> dn3. Before, we use multi-clients to send data
to dn1, but it does not improve so much. Because when dn1 send data to dn2, we
also should use multi-clients, but currently there is only one client(i.e.
RemoteStream#out) send data from dn1 to dn2.
2. There are only about 10 threads to do DataStreamManagement#read, dn1 send
data to dn2 maybe slowed by only one client, then DataStreamManagement#read
also maybe slow, so we should make DataStreamManagement#read async model.
I will submit 2 PR to adress #1 and #2.
> 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, image-2021-03-10-11-23-19-681.png,
> image-2021-03-10-11-23-42-092.png, screenshot-1.png, screenshot-2.png,
> streaming.svg
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
