I was thrown for a loop by a Transcender test prep question the other day.
After reading in the MSDN Library, I still didn't really grasp the just of
the question.  I was wondering if someone could give me a little more
explanation.

The question was as follows:
You are the system architect for a client-server application that will
support more than 10,000 users.  DCOM objects for your application will be
created on servers around the world.  You want to provide load balancing,
but normal dynamic load balancing is inappropriate because client
applications will remain connected to the server object for extended periods
of time.  How would you implement load balancing on a method-by-method
basis?

The answer is:
Make the client and server component stateless.

Any help is much appreciated.

Regards,

Travis Spencer
Portland, OR. USA

P.S.  To all those who hang out in the MCSD certification group, sorry for
cross posting.  My test is in two days, so please forgive me for wanting to
do all I can to find answers ASAP.

if they say that the server objects should be stateless, this is bullshit
too

these objects should not contain client specific state and the client cannot
expect that his state will survive method calls

i'm really sick of the word stateless... a to simple explanation of a really
difficult problem