You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mina.apache.org by "Trustin Lee (JIRA)" <ji...@apache.org> on 2007/11/05 02:53:50 UTC

[jira] Resolved: (DIRMINA-144) Traffic shaping filter

     [ https://issues.apache.org/jira/browse/DIRMINA-144?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Trustin Lee resolved DIRMINA-144.
---------------------------------

    Resolution: Fixed
      Assignee: Trustin Lee

I've just added TrafficShapingFilter, which works on a per-session basis.  It seems to work pretty well so far.

I have no idea on how to implement global or per-service traffic shaping yet.  Probably we could resolve it in other newly created issue.

> Traffic shaping filter
> ----------------------
>
>                 Key: DIRMINA-144
>                 URL: https://issues.apache.org/jira/browse/DIRMINA-144
>             Project: MINA
>          Issue Type: New Feature
>          Components: Filter
>            Reporter: Niklas Therning
>            Assignee: Trustin Lee
>            Priority: Minor
>             Fix For: 2.0.0-M1
>
>
> MINA now has basic support for suspending/resuming reads and writes on an IoSession for all transport types. We should use this feature to implement a traffic shaping filter which lets the user configure bandwidth limitations. Both globally and for individual sessions. Here are some configuration options I can come to think of:
> * Globally specify a maximum data rate for individual sessions (e.g. max 10 kb/s for all sessions).
> * Specify a maximum data rate on a per session basis (e.g. max 1 Mb/s for a particular session).
> * Specify a maximum upper limit on the bandwidth used by all sessions in total (e.g. all sessions share a total of 10 Mb/s).
> * Configure different traffic classes with different limits to allow for high/low priority traffic and partition sessions into these priority classes.
> I'm not sure how advanced it would have to be so please share your thoughts. Maybe point 1 and 2 would suffice? I think 3 and 4 add a lot of complexity since they need some kind of scheduling algorithm to provide fairness between sessions.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.