You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mina.apache.org by Julien Vermillard <jv...@archean.fr> on 2006/04/13 09:02:53 UTC

stupid latency test

Hi,

I had a discussion with Trustin about latency and I wrote a little test
(based of reserver example). The values looks pretty low (1ms). I wonder
if my test is correct.

Julien

Re: stupid latency test

Posted by Julien Vermillard <jv...@archean.fr>.
Le jeudi 13 avril 2006 à 16:11 +0900, Trustin Lee a écrit :
> On 4/13/06, Julien Vermillard <jv...@archean.fr> wrote:
>         Hi,
>         
>         I had a discussion with Trustin about latency and I wrote a
>         little test
>         (based of reserver example). The values looks pretty low
>         (1ms). I wonder
>         if my test is correct.
> 
> It seems like you didn't set any thread model.  You need to set the
> thread model to ThreadModel.MANUAL so it run in a single thread.
> 
> And is there any reason you spawn a new thread?  You could just store
> the current nanotime as a session attribute and then calculate the
> latency when messageReceived is invoked later.  It will decrease
> latency much more. 
> 
> BTW 1 ms latency sounds very good.

Yep I wanted ot test with default thread model config. I'll try with
ThreadPool 


I use to spawn a thread because in some case calling session.write(); in
the IoHandler used to create deadlocks. I think it's fixed in MINA now,
but it's an old habbit :)

Well I think Java precision is too low for mesure the real latency.
For me it's look low enougth :)

Julien


Re: stupid latency test

Posted by Trustin Lee <tr...@gmail.com>.
On 4/13/06, Julien Vermillard <jv...@archean.fr> wrote:
>
> Hi,
>
> I had a discussion with Trustin about latency and I wrote a little test
> (based of reserver example). The values looks pretty low (1ms). I wonder
> if my test is correct.


It seems like you didn't set any thread model.  You need to set the thread
model to ThreadModel.MANUAL so it run in a single thread.

And is there any reason you spawn a new thread?  You could just store the
current nanotime as a session attribute and then calculate the latency when
messageReceived is invoked later.  It will decrease latency much more.

BTW 1 ms latency sounds very good.

HTH,
Trustin
--
what we call human nature is actually human habit
--
http://gleamynode.net/
--
PGP key fingerprints:
* E167 E6AF E73A CBCE EE41  4A29 544D DE48 FE95 4E7E
* B693 628E 6047 4F8F CFA4  455E 1C62 A7DC 0255 ECA6