A traffic generator implementation framework for packet switched networks

Traffic generation for packet switched networks is a process in which we ar tificially inject packet streams into a target network. Traffic generation makes it possible to test and study the performance of network systems and their applications without incurring the substantial cost of the equipment and human testers that would otherwise be required. An effective network tr affic generator must provide scalability, emulation accuracy, validation ab ility, and flexibility. Unfortunately, no existing commercial packet switch ed network traffic generator offers all these features. In this paper, we d evelop a framework that provides an implementation platform for a generic t raffic generator. Upon the platform, a user can create a variety of service s, such as workload emulation, performance probing, and traffic profiling. The framework employs extended finite state machine (EFSM)-based modeling a nd an event-driven paradigm; these allow a traffic generator based on the f ramework to provide the features listed above.