Abstract: In order to enhance network reliability in avionics when there is a shortage of network equipment due to malfunction, two performance degradation constraints are defined. Both degradation constraints employ three arguments to accurately describe the number and distribution of message failures or successes during transmission. New concepts, such as dynamic failure, critical function and minimal guaranteed future sequence, are defined to analyze the performance degradation constraints: the value of critical function can predict whether a dynamic failure will occur if the transmission of the next message fails; the minimal guaranteed future sequence can minimize the number of successful transmissions without the occurrence of dynamic failure. Two strategies are proposed to implement the performance degradation constraints in avionic networks: static filtration by message source and dynamic arbitration by network. A double-layer priority scheduling algorithm is proposed, which is based on performance degradation constraints and used in the strategy of dynamic arbitration by network. The algorithm is able to avert dynamic failure by means of the predictability of critical function. The schedulability of the algorithm is also analyzed. The analysis and simulation proves that the network reliability in avionics is improved by implementing the performance degradation constraints.

Key words: avionics, reliability, network performance, degradation, schedulability analysis, dynamic failure