Journals Proceedings

Abstract

The paper presents experimental results on the optimal height of the EFSM equivalent for deriving high quality tests for telecommunication protocols. The EFSM model is a widely used finite state model, and its l-equivalent is an FSM which behavior is equivalent to the initial EFSM for each input sequence of the length l or less. The l-equivalent is known to be more suitable for test derivation than the initial EFSM and thus, it is often used as the specification EFSM for this purpose. In this paper, test derivation the l-equivalent has been derived for the following protocols: POP 3, SMTP, TIME, DCCP, TCP. EFSM models have been extracted from the corresponding RFC specifications, and mutation testing techniques have been utilized in order to estimate the fault coverage of the test suites which correspond to a transition tour of a corresponding l-equivalent. Experimental results clearly show that even when the height l of the l-equivalent is low, and moreover, equals two, many functional faults can be detected in the protocol implementation.