Re: OpenJMS Scalability

Take a look at JGroups ... this is framework that has really impressed 
me of late.  I have just completed a project where i created my own Grid 
like network, where i have tasks and i find the next available worker to 
run them.

I am using JGroups as the transport mechanism, which is configured to 
use MultiCast.  JGroups can use JMS as its transport mechanism as well 
if you want.


R.F. Pels wrote:
> On Friday 16 September 2005 20.06, Steve Angelovich wrote:
> 
>> We have a distributed java application that does a lot of scientific
>> computations on a Linux cluster.  In our environment we may start a job
>> that runs on 16 or 32 nodes and may run for several days before
>> completing.  The decomposition is pretty simple so there typically isn't
>> a lot of node to node communication required once the initialization
>> gets completed.  However there are  cases where a node will query all
>> the other nodes.  A simple example would be to find out their status.
> 
> I'm unsure if 1000+ clients scales well. But wouldn't you in this case better 
> consider multicasting? As a matter of fact, I'm investigating into Mule. It 
> can use MOM if you want to, but you're not restricted to it.



-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. 
Download it for free - -and be entered to win a 42" plasma tv or your very
own Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php