Efficient network management and security in 5G enabled internet of things using deep learning algorithms

Abstract

The rise of fifth generation (5G) networks and the proliferation of internet-of-things (IoT) devices have created new opportunities for innovation and increased connectivity. However, this growth has also brought forth several challenges related to network management and security. Based on the review of literature it has been identified that majority of existing research work are limited to either addressing the network management issue or security concerns. In this paper, the proposed work has presented an integrated framework to address both network management and security concerns in 5G internet-of-things (IoT) network using a deep learning algorithm. Firstly, a joint approach of attention mechanism and long short-term memory (LSTM) model is proposed to forecast network traffic and optimization of network resources in a, service-based and user-oriented manner. The second contribution is development of reliable network attack detection system using autoencoder mechanism. Finally, a contextual model of 5G-IoT is discussed to demonstrate the scope of the proposed models quantifying the network behavior to drive predictive decision making in network resources and attack detection with performance guarantees. The experiments are conducted with respect to various statistical error analysis and other performance indicators to assess prediction capability of both traffic forecasting and attack detection model.