Attack Impact Analysis of Service Function Chain in Aviation Communication Network

Abstract

With the integration of technologies such as cloud computing, edge computing, and network function virtualization (NFV) into aviation communication networks, the cost of communication services has decreased while flexibility has increased, but this has also expanded the attack surface of the network. Currently, there is no research on the impact of attacks on service function chains (SFCs) in aviation communication networks. In order to address this issue, a formal modeling of SFC deployment is first conducted, followed by an analysis of the recovery process of SFC after the network has been attacked. Second, a case analysis is performed to refine the SFC attack impact analysis model, which introduces six typical attack strategies and four SFC recovery methods proposed in the literature. Finally, experiments were conducted to investigate the impact of different attack strategies on the availability, quality of service, and security of SFC in aviation communication networks and to compare the robustness of different SFC recovery methods in responding to these attacks. The experiments found that betweenness-based attacks have the most significant impact on the SFC failure rate, the number of virtual link (VL) redirections, and the average increase rate of coexistence risk, increasing them by 170.85%, 108.39%, and 272.94%, respectively. Load-based attacks have the greatest effect on the number of virtual network function (VNF) migrations, increasing it by 120.65%. Degree-based attacks have the largest impact on the average latency variation of the SFC, increasing it by 170.20%. Additionally, it was found that among the four SFC recovery methods proposed in the literature, the recovery strategy based on evolutionary algorithms shows the best robustness against attacks in terms of SFC failure rate and average delay variation, while the recovery strategy based on greedy algorithms performs best regarding the number of VNF migrations, VL redirections, and average increase rate of coexistence risk.