Re: [xmlblaster] Testing Protocols

[Zied ABID <zied.abid at it-sudparis.eu>]

Dear Mohannad,

If I understand, you want just use xmlBlaster to publish and receive 
message, so you should take a look on javaClients HelloWorld in
$XMLBLASTER_HOME/demo/javaClients specially : HelloWorldPublish and 
HelloWorldSubscribe

$XMLBLASTER_HOME/demo/HelloWorld6 is helpful too to have an idea on 
protocol's configuration


I'm begging also in xmlBlaster , so I hope that is helpful


Zied


Mohannad a écrit :
> thanks Marcel,
>
> I need the XMLBlaster for just publish messages and the clients simply 
> will receive them. So by changing the code of Java file in the client 
> side (will lead somehow to different seniors), and then we can take 
> some approximate measurements.
>
> If this possible at least to build some approximation and not 
> necessary to be very accurate, I will be appreciate it if you could 
> help us to make it by this Middleware package.
>
> Thanks again,
>
> Mohannad
>
> On Wed, Apr 16, 2008 at 11:00 AM, Marcel Ruff <mr at marcelruff.info 
> <mailto:mr at marcelruff.info>> wrote:
>
>     Mohannad wrote:
>
>         Dear Marcel,
>
>         I'm new here for using XMLBlaster. I am working with my
>         college in my university on project which depends totally on
>         XMLBlaster package. We need the package for testing several
>         protocols on different operating systems (different objects
>         also) by changing the loads that sent. This will help to
>         decide which protocol is better than other in specific loads.
>
>         My doctor in my university suggested to use XMLBlaster since
>         it has powerful capability to monitor the real time and then
>         calculate statistics (e.g number of messages sent in specific
>         time period)
>         I tried to figure out that capabilities by using jconsole, but
>         I could not find these real time monitoring for my project's goal.
>
>         Could you help me to find a way to accomplish our project goal?
>
>     Hi Mohannad,
>
>     you plan a scientific task.
>
>     To do such performance measurements you need to create an accurate
>     experimental setup
>     and precisely define the measurement points with minimal 'noise'
>     (cross effects etc) in between.
>
>     This needs to be coded depending on what your goals are
>     and can't be generally done by a middleware.
>
>     I wish you success,
>     Marcel
>
>
>         Thanks Marcel,
>
>         Mohannad
>
>
>
>
>     -- 
>     Marcel Ruff
>     http://www.xmlBlaster.org
>     http://watchee.net
>     Phone: +49 7551 309371
>
>
>
>
> --
> Mo|-|annad
> -وما توفيقي الا بالله -