New Approach for EFSM-Based Passive Testing of Web Services

Fault management, including fault detection and location, is an important task in management of Web Services. Fault detection can be performed through testing, which can be active or passive. Based on passive observation of interactions between a Web Service and its client, a passive tester tries to detect possible misbehaviors in requests and/or responses. Passive observation is performed in two steps: passive homing and fault detection. In FSM-based observers, the homing consists of state recognition. However, it consists of state recognition and variables initialization in EFSM-based observers. In this paper, we present a novel approach to speed up homing of EFSM-based observers designed for observation of Web Services. Our approach is based on combining observed events and backward walks in the EFSM model to recognize states and appropriately initialize variables. We present different algorithms and illustrate the procedure through an example where faults would not be detected unless backward walks are considered.