directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Trustin Lee <trus...@gmail.com>
Subject Re: [mina] Vinod's perf stats
Date Sat, 07 May 2005 05:50:05 GMT
Hi,

> On 5/7/05, Alex Karasulu <aok123@bellsouth.net> wrote:
> > Hey guys,
> >
> > IMHO the main page for MINA could have some of those stats for the
> > concurrency and other stats that Vinod generously reported to the list.
> > It would really show how much of a better alternative MINA really is in
> > terms of performace in comparison to other frameworks.  Just wanted to
> > get the idea out what do you thnk Trustin?

Sounds good, but we'll have to refine the benchmark framework first
that will take some time.

2005/5/7, Vinod Panicker <vinod.p@gmail.com>:
> Those stats really didnt mean anything concrete.  Its a great idea to
> put performance stats on the page though.  I'd advise that we hold on
> till we do tests related to throughput as well.  Also, the NIO impl in
> the JVM has serious performance issues.  CPU and memory consumption
> are way too inflated.  We have to figure out a way around this to show
> MINA's real capabilities.

You're right.  I don't think we can make any hot performance chart
till the release.  Let's do this after we release 0.7.

BTW do you guys know that MINA can support non-NIO?  We can simply
implement stream-based I/O layer implementation that provides one
thread per one socket thread model so that users choose what they
prefer.

Trustin
-- 
what we call human nature is actually human habit
--
http://gleamynode.net/

Mime
View raw message