Consider a SINGLE server system with customers that arrive to the station according to a Poisson process. The server starts offering service in "normal" mode, but jumps into "fast" mode after completing a service if sees more than m customers waiting in the queue, remaining "fast" until the system empties, going back to "normal" mode, and so on. The service times are drawn from a stream {tfast} or {tnormal}. Present an event graph, including parameters and state variables, that could be employed to estimate the fraction of time the system operates in fast mode. Consider a SINGLE server system with customers that arrive to the station according to a Poisson process. The server starts offering service in "normal" mode, but jumps into "fast" mode after completing a service if sees more than m customers waiting in the queue, remaining "fast" until the system empties, going back to "normal" mode, and so on. The service times are drawn from a stream {tfast} or {tnormal}. Present an event graph, including parameters and state variables, that could be employed to estimate the fraction of time the system operates in fast mode