[ 
https://issues.apache.org/jira/browse/MESOS-1330?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14148257#comment-14148257
 ] 

Joris Van Remoortere commented on MESOS-1330:
---------------------------------------------

[~vinodkone]: I added a benchmark test which exercises libprocess in a many-1 
fashion. It forks in order to ensure there is no short-circuiting of the 
process-manager dispatch.
The benchmark test can be easily changed by the "num_iter", "num_threads", and 
"queue_depth" variables at the top.
I ran it on the current apache mesos master, as well as on my prototype. The 
result differences are within the variance of the test itself:
https://github.com/mesosphere/mesos/commits/abstract-event-manager
Master / Prototype (RPCs / second; 120K sample)
[42172] / [43116]
[42890] / [42812]
[43126] / [42950]
[42861] / [42829]

avg:
[42762] / [42926]

The key take-away is that currently both implementations are bottle-necked by 
contention rather than CPU. This is something I hope to address as a separate 
performance issue.

> Introduce connection/transport abstraction to stout
> ---------------------------------------------------
>
>                 Key: MESOS-1330
>                 URL: https://issues.apache.org/jira/browse/MESOS-1330
>             Project: Mesos
>          Issue Type: Improvement
>          Components: general, libprocess
>            Reporter: Niklas Quarfot Nielsen
>              Labels: libprocess, network
>
> I think it makes sense to think in terms of different low or middle layer 
> transports (which can accommodate channels like SSL). We could capture 
> connection life-cycles and network send/receive primitives in a much explicit 
> manner than currently in libprocess.
> I have a proof of concept transport / connection abstraction ready and which 
> we can use to iterate a design.
> Notably, there are opportunities to change the current SocketManager/Socket 
> abstractions to explicit ConnectionManager/Connection, which allow several 
> and composeable communication layers.
> I am proposing to own this ticket and am looking for a shepherd to 
> (thoroughly) go over design considerations before jumping into an actual 
> implementation.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to