A Model to Calculate Exact End-to-End Delay of Sporadic Flows on AFDX Network Using Mathematical Programming

Abstract

Avionics Full DupleX (AFDX) Switched Ethernet technology provides a deterministic network to avionics context. Guarantees on worst-case end-to-end (ETE) communication delays are required for certification purposes. This paper identifies the worst-case ETE analysis as an optimization problem and proposes a mathematical programming approach to calculate an exact worst-case delay and the corresponding scenario. It allows to significantly increase the size of the configuration for which an exact worst-case can be determined. The approach considers a set of sporadic flows with no assumption concerning the arrival time of packets.