A SDN-based Traffic Bandwidth Allocation Method for Time Sensitive Networking in Avionics

Abstract

The Integrated Modular Avionics (IMA) network systems confront bandwidth resource sharing challenge constantly, with the objective to obey strict timing requirements. In traditional avionics network, the deterministic QoS guarantees are maintained by precise offline configuration during the design period. However, considering the increasing flexibility and variable scale of avionics systems in recent years, it is inevitable to develop a new network framework that can adapt to scalability. Thus, a Software Defined Networking (SDN)-based avionics network is explored to reduce the complexity and integration overheads for upgrading, which must meet bandwidth and end-to-end latency requirements both. SDN possesses the advantages with global visibility and flexibility management of flows for safety critical applications, but it could not support real-time guarantee for traffic transmission in the network. Nevertheless, the centralized SDN can be an effective implement to combine SDN and bandwidth resource allocation for Time Sensitive Networking (TSN) in avionics network. And it is revealed that the Time Aware Shaper (TAS) and Credit Based Shaper (CBS) play key roles in bandwidth allocation for TSN to obtain deterministic traffic transmission behaviors, which also shows great interests to automotive and industrial communications. In this paper, we first propose a SDN-based framework for traffic bandwidth allocation in avionics network, which can significantly enhance bandwidth reservation management during real-time network communication. Then, we introduce a heuristic algorithm to resolve traffic bandwidth allocation problem, and derive multiple scheduling constraints to support traffic transmission with low and bounded latency and further ensure safety-critical communication. Experiments are carried out to demonstrate our method, and we also discuss trade-offs to optimize the required calculation time, thus can be flexible adopted in safety-critical and time-sensitive applications.