Robust Priority Assignment applied to AFDX networks

Abstract

The priority assignment of flows on an AFDX network impacts directly the end-to-end delay of all messages. This paper proposes the use of Robust Priority Assignment (RPA) in order to increase the additional interference that the most vulnerable virtual links of a AFDX network can support by decreasing their worst-case end-to-end delays through a better priority assignment. A comparison is performed with other stateof-the-art techniques used to assign priorities. The results shows that the use of RPA increases the network robustness in 307 in a sample configuration.