[Spread-users] Spread Queuing question

[Alan]

I am just now learning about Spread to implement a very high load event 
system.  I am struggling to understand how I can leverage Spread to 
distribute my load.  My current plan is to publish to a Spread Cluster 
and then have Event Daemons recording those events to the DB.  I found a 
previous article on Spread and Queuing at: 
http://commedia.cnds.jhu.edu/pipermail/spread-users/2005-November/002766.html, 
but I am failing to see how I can direct traffic to a single client, 
ensure its delivery, and let other clients pick up other messages.

I want to make sure I can guarantee delivery (although I don't care 
about order), and support scaling out the logging of events.  I cannot 
slow down processing of the main application, so pushing to Spread makes 
sense to have a comfort level around the message being caught.  Pulling 
from Spread to record the event is where I am concerned.  I suppose I 
don't completely understand how scaling out to more Spread Daemons on 
more machines can help improve performance.

Any help would be appreciated.

Thanks,

Alan