Detection of evil flies: securing air-ground aviation communication

Abstract

The aviation community is employing various air traffic control and mobile communication technologies, such as ubiquitous data links, wireless communication architectures and protocols. Recently, software-defined networking (SDN) based architectures (i.e., cockpit network communications environment testing (COMET)) have been proposed for Air-Ground communication. However, an evil can break the communication between a pilot and air traffic control, resulting in a hazardous (or life-threatening) situation up in the air or failure of ground equipment. This paper proposes an efficient evil detection and prevention mechanism (called DoEF) for the COMET architecture. The proposed DoEF utilizes a deep learning-based approach, i.e., long-short term memory (LSTM), to detect the evil flies and provide possible countermeasures. Our preliminary results show that the proposed scheme reduces the detection time and increases the detection accuracy of distributed denial of service (DDoS) attacks for the aviation network.