Research On Fault Diagnosis Of Protocol Conformance Testing

The purpose of protocol testing is to ensure the protocol implementation executed steadily and reliably according to the protocol specification. Recent advances in communication technology have made protocol testing more and more important. Conformance testing is the basis of other protocol testings. There are two kinds of conformance testing methods: active testing and passive testing. Active testing method is that testing system actively controls the input sequences to IUT (Implementation under Test), then based on IUT's response to judge whether IUT is consistent with the specification. The advantage of active testing is that test sequences can be carefully chosen at need. Passive testing method is that testing system doesn't communicate with IUT, just passively collecting the mutual messages between IUT and other systems, then analyzing these messages to find faults of IUT. As not interfereing with the normal running of network systems, passive testing method is the best method to find the potential faults on the real network. After a fault has been detected, an important problem how to diagnose the fault occurs.This paper focuses research on fault diagnosis and includes three major parts: 1. Fault Diagnosis Based on Finite State Machine Model in Active Testing After a fault has been detected, it arises the problem that how to diagnose the fault. In this paper, we propose a new fault diagnosis algorithm after analyzing the deficiency of an existing fault diagnosis algorithm. The diagnosis algorithm makes full use of transitions confirmed to be correct and head states set of the next observed input/output pair corresponding to the symptom transition, and guarantees the diagnosis of any single fault in an FSM (Finite State Machine) . In this paper, the comparison result of time complexity with other existing fault diagnosis algorithms is given out and shows that the new algorithm is more efficient.Then, we propose another fault diagnosis algorithm for UIO-based test sequences, which utilizes as much information of UIO test sequences as possible. The algorithm can decrease the complexity significantly and guarantee the diagnosis of any single...