Efficient handover protocol for 5G and beyond networks

Abstract

The fifth generation (5G) and beyond 5G (B5G) networks offer ultra-low latencies, higher reliability, scalability, data rates and capacities to support applications such as vehicular communications, internet of everything (IoE) and device to device (D2D) communication. In spite of these excellent features, user privacy, resource management and handover authentications present some challenges. To facilitate seamless connectivity in 5G and B5G networks, numerous machine learning schemes have been developed to facilitate target cell selection based on parameters such as signal strength and signal to noise ratio (SNR). However, most of these approaches concentrate on performance enhancements, ignoring security and privacy issues. On their part, majority of the conventional handover authentication schemes exhibit long latencies which contravenes 5G and B5G requirements. Moreover, the base stations in these networks have very small footprints and hence require the deployment of numerous base stations within the coverage area. This serves to compound performance, security and privacy issues due to the resulting frequent handovers. In this paper, a multilayer neural network (MLNN) privacy and security preservation protocol is presented. To facilitate target cell selection, parameters that took user satisfaction, network, user equipment (UE) and service requirements into consideration were deployed so as to enhance both quality of service (QoS) and quality of experience (QoE) during and after handover. For handover security, timestamps, ephemerals and random nonces were deployed during handover authentication to offer both security and privacy. Formal security analysis using Burrows-Abadi-Needham (BAN) showed that the proposed protocol offered strong mutual authentication among the communicating entities. On the other hand, informal security analysis showed that the proposed protocol offers perfect forward key secrecy and is robust against attacks such as impersonation and packet replays. In addition, performance evaluation showed that it has the lowest communication costs and average computation overheads. Moreover, it exhibited a 27.1% increase in handover success rate, and a 24.1% reduction in ping pong rate.