Integrated reliability theory and evaluation methodology of AFDX

This paper focuses on the fault-tolerant operation mechanism and failure mechanism of Avionics Full Duplex Switched Ethernet (AFDX). With the strong correlation among airborne systems, fault tolerance mechanism of full duplex communication and dynamic adjustment strategy, this paper presents an integrated reliability theory and evaluation methodology based on service performance and fault tolerant degradation. In order to describe the correlation among hardware, software and protocol, stochastic Petri network is utilized to establish their reliability model based on failure coverage rate. Directing to the sudden traffic variance under failures, this paper presents an optimal threshold traffic model based on sudden traffic and steady traffic, in which the particle swarm algorithm will be used to get the optimization threshold. With the control theory, this paper will design the nonlinear robust resources controller to adjust the buffer size dynamically. Selecting the time delay as service performance, G/M/1 queuing model is explicated to analyze the traffic shaping and the network calculus algorithm is used to get the maximum network time delay. Combing the time delay of traffic shaping and adjustment, the service-oriented integrated reliability model of AFDX is presented based on time delay and structure reliability to realize the reliability evaluation for airborne network - AFDX of commercial airplane.