My experiment: 1) 2 AppDomains within one process: AppDomain A sends AppDomain B 100 messages via Remoting or Socket AppDomain B sends AppDomain A 100 messages via Remoting or Socket
2) 2 Processes: Process A send Process B 100 messages via Remoting or Socket Process B send Process A 100 messages via Remoting or Socket For 1 and 2, I measure the delta difference. And the delta difference I got is almost not significant. But I am not sure if AppDomain will show improvement over process if the number of process is increasing: ie, write 30 AppDomain within one process vs write 30 processes. The CPU spent on process context switch will then show the delta.
