Energy-efficient dynamic homomorphic security scheme for fog computing in IoT networks

Abstract

Recently, there is an exponential increase in the multimedia and other data over the Internet of Things (IoT). This data is generally send to the cloud for processing and storage. The fog layer in-between readily bridges communication among the IoT devices and the cloud. It delivers services efficiently by computing and analyzing various multimedia information generated by the IoT devices residing on the sensors. However, provision of effective security and energy are critical challenges. The purpose of this work is to enhance the secure transfer of information like multimedia. This scheme uses Message Queue Telemetry Transport (MQTT) protocol over SSL/TLS. Since MQTT is vulnerable to eavesdropping, the Elliptic curve-ElGamal cryptography algorithm is introduced which lends a homomorphic factor thereby mitigating man-in-the-middle attack. The dynamic key change and proportional offloading of data as proposed in the current research work helps to preserve node energy by selectively transferring data to the cloud and the fog according to the data topic. The results depict that the system security and lifetime can be improved in comparison to the existing protocols.

Introduction

In the recent years, IoT has been evolved to a paradigm which makes it possible to interconnect the physical objects and humans in smart cities/homes etc. This has been possible due to the advancement in various communication and networking technologies [1]. Moreover, multimedia IoT systems have taken a greater leap, and have been widely used in surveillance,industrial automation [2], [3], etc. Although IoT offers numerous benefits, it also faces challenges in this field. The four biggest issues looming this sector are privacy [4], [5], [6], cybersecurity [7], [8], energy consumption [9], [10] and efficient routing protocols [11], [12].

To address these issues, cloud computing has been introduced in IoT, and data is transferred to this platform for analysis and processing. The concept of device-to-device(D2D) communication-aided fog computing [13] has evolved recently. It authenticates proximal devices without the use of a centralized server. It has also proved fruitful in coping with security and privacy issues [14]. To deal with latency and bandwidth issues in the cloud, fog computing and mobile edge computing (MEC) paradigms have emerged in the recent years. This work introduces a similar hierarchical architecture with intermediate fog layer, as shown in Fig. 1, depicting the flow of information. The fog layer acts as a gateway by filtering the data, processing it and sending the heavier computations to the cloud layer. In this scheme, the data is divided according to MQTT topics [15]. If the message topic needs fewer computations, it is directed to the fog nodes else it is offloaded to the cloud data centers for heavier computations. The major contributions and issues this work resolves are as follows:

• Firstly, Proportional data offloading is proposed in this work which decreases energy consumption, a major concern of an IoT framework.

• Secondly, this work aims to increase security by using hybrid asymmetric cryptography algorithm namely ECC-ElGamal, a homomorphic encryption [16] technique. Elliptic Curve Digital Signature Algorithm (ECDSA) is implemented in the application layer to maintain proof of identities and data integrity.

• Dynamic Key Change is implemented to mitigate eavesdropping and man-in-the-middle (MITM) attacks. The dynamic key change helps to conserve system energy and protects the system from the security attacks by using different curves based on the node energy level.

• The simulation has been performed in this environment and the results depict that the proposed algorithm improvises the energy consumption and enhances security.

The dynamic key change and proportional offloading of the data are the main context parameters that are related to node battery and security. Thus, the framework is designed to maintain confidentiality, security and integrity in the system.

The three-tier fog computing system as shown in Fig. 2 comprises the following three entities: the cloud data centers, the fog server, and the sensor/actuator module with inter and cross-layer communication. The heterogeneous data from the sensor devices reach a central gateway or the base station which processes the data before sending it to the remote servers through MQTT.

The remainder of the paper gives an insight into the proposed framework and methodologies used. In Section 2, various security and energy related literature is reviewed. In Section 2.3, motivation and contributions of the work are discussed. Section 3, discusses the different parameters used in our work. In Section 4, the proposed work is introduced which includes the system model, followed by the security algorithm under which dynamic key change according to node battery and homomorphic encryption is discussed. In Section 5, security and energy analysis proves that with the right selection of parameters, power consumption can be substantially improved and it can mitigate security issues such as MITM attacks. In Section 6, the performance of the proposed system is evaluated. Finally, conclusions are drawn and future scope of this work is discussed in Section 7.