[
https://issues.apache.org/jira/browse/DIRMINA-332?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Gaston Dombiak updated DIRMINA-332:
-----------------------------------
Attachment: changes.patch
> Improve performance of StatCollector and change obtained statistics
> -------------------------------------------------------------------
>
> Key: DIRMINA-332
> URL: https://issues.apache.org/jira/browse/DIRMINA-332
> Project: MINA
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.0.1
> Reporter: Gaston Dombiak
> Fix For: 1.1
>
> Attachments: changes.patch
>
>
> StatCollector#addSession and StatCollector#removeSession are blocking the
> entire object. Therefore under heavy load of new sessions or remove of
> sessions we will have a bottleneck. Instead we can take advantage of "new"
> Java 5 concurrent util classes.
> Throughput is incorrectly being estimated. It is missing packets/bytes and
> when dealing with a very big number of sessions the rate will be off. Instead
> of keeping stats of throughput we can just stats of total traffic.So stats
> will change from --> to
> MsgWrittenThroughput --> totalMsgWritten
> MsgReadThroughput --> totalMsgRead
> BytesWrittenThroughput --> totalBytesWritten
> BytesReadThroughput --> totalBytesRead
> The rate can then be estimated by the client of this class. We can add 2 new
> stats that are very helpful (at least to me):
> 1) scheduled writes: Number of current scheduled writes for all existing
> sessions.
> 2) queueved events: Number of current queued events for all existing sessions.
> Stat #1 can always be estimated. However stat #2 can only be estimated when
> the ExecutorFilter is present in the filter chain.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
