Hi all- 

SDK question - what is the recommended method to manage an instance of PipelinePilotServer?
I anticipate creating several jobs within a short period of time (a couple of minutes) and I'd like to reuse a previous instance of PipelinePilotServer to speed the creation of those jobs within a single thread in my project (it's a Workbook-based extension). However, I don't want to assume that the connection will persist for my users' entire session, since if it doesn't later job submissions will fail.

I guess there are really a couple of additional questions here:
- when one creates an instance of PipelinePilotServer, how long should one assume it's valid?
- if there is a timeout period on the server, is it based on the overall connection time or the time elapsed since the last job was run?
- is there any performance risk of keeping hundreds of these instances open across our enterprise application (that's the number of simultaneous users)? 
- are these answers different if one is using a load-balanced PP server environment?

FWIW we have deployed 9.5 SP1 in a load-balanced production environment (3 servers, WS2008R2); in the future we plan on moving our Foundation infrastructure to Linux.

Thanks-
Marty