Handling redundant and additional states in protocol testing

Abstract

This paper addresses the problem of conformance testing of protocols modeled by FSMs with redundant states. Redundant states appear in an FSM which may be nonminimal or nonconnected. The existing test derivation methods usually are not directly applicable to these machines. In this paper, we show that they can be adjusted to cover this class of FSMs and that the traditional assumption on the minimality of machines is not necessary. Another problem with redundant states is that they can cause the appearance of additional states in protocol implementations whose guaranteed detection requires tests of an exponential length. This paper proposes techniques for deriving tests for FSMs with redundant or additional states such that a high fault coverage is achieved while maintaining an acceptable test suite length. The effectiveness of the proposed methods has been evaluated in an experimental way using a benchmark protocol.

Testing timed automata

“…A remarkable property of our method is that, unlike most testing approaches for Mealy machines (with some exceptions, e.g., [27]), we do not assume minimality of the speciÿcation and implementation automata. Nevertheless, for reasons of e ciency it is of course desirable to work with minimal automata.…”

Exploiting Symmetry in Protocol Testing

“…The main idea is that test derivation is not based on the entire specification autom aton, but only on a kernel of it. A technical detail here is th at we do not require Mealy machines to be minimal (as already observed by [20] for the setting w ithout symmetry). □ In Chow's paper, conformance is defined as the existence of an isomorphism between spec ification and implementation.…”

Towards testable communication software

“…In this case, we may assume that this state has an infinite transfer sequence. Note that FSMs with unreachable states usually require lengthy test suites [PHK95]. A state cover constitutes a skeleton of test suites with guaranteed or full fault coverage in the sense that every transition from every state specified in the machine should be tested by such a test suite.…”