You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ratis.apache.org by "runzhiwang (Jira)" <ji...@apache.org> on 2020/12/07 00:55:00 UTC

[jira] [Comment Edited] (RATIS-1209) Compare the performance between DataStreamApi and AsyncApi

    [ https://issues.apache.org/jira/browse/RATIS-1209?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17244881#comment-17244881 ] 

runzhiwang edited comment on RATIS-1209 at 12/7/20, 12:54 AM:
--------------------------------------------------------------

[~szetszwo] Hi, I test the  https://github.com/apache/incubator-ratis/pull/327. But result is same, DataStreamApi cost 18402ms, and AsyncApi cost 17748ms. I also attach the perf result of primary and peer when use DataStreamApi , hopes it can help.


Command of DataStreamApi
 ${BIN}/client.sh filestore datastream --size 100000000 --numFiles 10 --bufferSize 4000000 --type DirectByteBuffer --peers ${PEERS}
Command of AsyncApi
${BIN}/client.sh filestore loadgen --size 100000000 --numFiles 10 --bufferSize 4000000 --peers ${PEERS}


was (Author: yjxxtd):
[~szetszwo] Hi, I test the  https://github.com/apache/incubator-ratis/pull/327. But result is same, DataStreamApi cost 18402ms, and AsyncApi cost 17748ms. I also attach the perf result of primary and peer when use DataStreamApi , hopes it can help.

> Compare the performance between DataStreamApi and AsyncApi
> ----------------------------------------------------------
>
>                 Key: RATIS-1209
>                 URL: https://issues.apache.org/jira/browse/RATIS-1209
>             Project: Ratis
>          Issue Type: Sub-task
>            Reporter: runzhiwang
>            Assignee: Tsz-wo Sze
>            Priority: Major
>         Attachments: 20201207-084349.svg, 20201207-084353.svg
>
>          Time Spent: 2h 20m
>  Remaining Estimate: 0h
>
> Not sure what happened.  If 3 servers are on different machines, DataStreamApi with type DirectByteBuffer seems slower than AsyncApi.  If 3 servers are on same machine, DataStreamApi with type DirectByteBuffer is better than AsyncApi.
> DataStreamApi command:
> `${BIN}/client.sh filestore datastream --size 100000000 --numFiles 10 --bufferSize 4000000 --type DirectByteBuffer --peers ${PEERS}`
> AsyncApi command:
> `${BIN}/client.sh filestore loadgen --size 100000000 --numFiles 10 --bufferSize 4000000 --peers ${PEERS}`



--
This message was sent by Atlassian Jira
(v8.3.4#803005)