Deep Learning and Blockchain-Empowered Security Framework for Intelligent 5G-Enabled IoT

Shailendra Rathore,Jong Hyuk Park,Hangbae Chang

doi:10.1109/access.2021.3077069

Abstract

Recently, many IoT applications, such as smart transportation, healthcare, and virtual and augmented reality experiences, have emerged with fifth-generation (5G) technology to enhance the Quality of Service (QoS) and user experience. The revolution of 5G-enabled IoT supports distinct attributes, including lower latency, higher system capacity, high data rate, and energy saving. However, such revolution also delivers considerable increment in data generation that further leads to a major requirement of intelligent and effective data analytic operation across the network. Furthermore, data growth gives rise to data security and privacy concerns, such as breach and loss of sensitive data. The conventional data analytic and security methods do not meet the requirement of 5G-enabled IoT including its unique characteristic of low latency and high throughput. In this paper, we propose a Deep Learning (DL) and blockchain-empowered security framework for intelligent 5G-enabled IoT that leverages DL competency for intelligent data analysis operation and blockchain for data security. The framework’s hierarchical architecture wherein DL and blockchain operations emerge across the four layers of cloud, fog, edge, and user is presented. The framework is simulated and analyzed, employing various standard measures of latency, accuracy, and security to demonstrate its validity in practical applications.

Highlights

The recent development in communication and networking applications gives rise to a massive demand for a generation communication paradigm
Based on the required design principle, we propose a Deep Learning (DL) and blockchain-empowered security framework for intelligent 5G-enabled Internet of Things (IoT) that employs DL and blockchain’s capability to support intelligent data and security operations across the 5G-enabled IoT network
The data analytic task is carried out on a cloud layer that consists of the following major components: Raw Data Collection: The data analytic operation starts with the collection and management of raw data consisting of different types from diverse sources, including Social Networking Services (SNSs), mobile devices, IoT devices, wearable devices, and many more

Summary

INTRODUCTION

The recent development in communication and networking applications gives rise to a massive demand for a generation communication paradigm. State-of-the-art security solutions have emerged to provide secure services and data-intensive applications in traditional 4G-enabled IoT These solutions support either centralized approach, novel paradigms, architecture, or framework for efficient, secure data management (i.e., processing, analysis, and storing) in the cloud [9]–[11]. As a strong analytic tool, DL can deliver stateof-the-art accuracy and latency than the traditional machine learning approach, and it can be deployed to analyze massive data in 5G-enabled IoT. Such DL deployment can support the prediction of future event and detection of attacks and provide substantial information for content caching and placement in dynamic scenarios of 5G-enabled IoT [17], [18]. The performance was evaluated on an object detection application using various standard measures of latency, accuracy, and security

REQUIRED DESIGN PRINCIPLES

EXPERIMENTAL EVALUATION

Findings

CONCLUSION