[jira] [Comment Edited] (STORM-2359) Revising Message Timeouts

["Stig Rohde Døssing (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/STORM-2359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16723067#comment-16723067 ] 

Stig Rohde Døssing edited comment on STORM-2359 at 12/17/18 3:21 PM:
---------------------------------------------------------------------

There's a branch implementing this at [https://github.com/srdo/storm/tree/STORM-2359-experimentation.] I did a couple of runs of the ConstSpoutNullBoltTopology from storm-perf, as well as the ThroughputVsLatency topology.

I'm seeing an overhead of about 10% for the ConstSpoutNullBoltTopology, since the spout and ackers are the only components involved. For the ThroughputVsLatency topology, the overhead looks to be negligible. I've attached the raw numbers and some charts.

The tests were run locally, with the default configuration for Storm as well as the default configuration for ConstSpoutNullBoltTopology. For TVL I reused the setup from the Netty benchmarks: ./storm jar .\storm-loadgen-2.0.1-SNAPSHOT.jar org.apache.storm.loadgen.ThroughputVsLatency --rate 90000 --spouts 4 --splitters 4 --counters 4 --reporter 'tsv:test.csv' -c topology.workers=4 --test-time 5


was (Author: srdo):
There's a branch implementing this at [https://github.com/srdo/storm/tree/STORM-2359-experimentation.] I did a couple of runs of the ConstSpoutNullBoltTopology from storm-perf, as well as the ThroughputVsLatency topology.

I'm seeing an overhead of about 10% for the ConstSpoutNullBoltTopology, since the spout and ackers are the only components involved. For the ThroughputVsLatency topology, the overhead looks to be negligible. I've attached the raw numbers and some charts.

> Revising Message Timeouts
> -------------------------
>
>                 Key: STORM-2359
>                 URL: https://issues.apache.org/jira/browse/STORM-2359
>             Project: Apache Storm
>          Issue Type: Sub-task
>          Components: storm-core
>    Affects Versions: 2.0.0
>            Reporter: Roshan Naik
>            Priority: Major
>         Attachments: STORM-2359.ods
>
>
> A revised strategy for message timeouts is proposed here.
> Design Doc:
>  https://docs.google.com/document/d/1am1kO7Wmf17U_Vz5_uyBB2OuSsc4TZQWRvbRhX52n5w/edit?usp=sharing



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)