ws reliability notes

A concern from the group about interacting with service instances
was scalability: the group wanted to make sure it could handle 1000+ 
messages/second.

My description of a service instance was as of an identifiable 
individual request,
as I take it from the ASAP draft.
Perhaps my description of the instance as "flyweight" to clarify that 
it was not necessarily
a full-fledged running process was inappropriately suggesting the 
requests
be in memory as threads of execution, when in fact I was trying to 
convey it
was an identifiable piece of work.

Is this a fair description?
Are there scalability issues?

I was more concerned about the specification for the factory and 
creation payload
than scalability personally...
(Perhaps refactoring for use with the grid service factory concept)

Comments?