Cyber Physical Systems - Advances and Applications

A Comprehensive Analysis of Various Threat Detection and Prevention Techniques in IoT Environment

P.S. Pavithra¹, *, P. Durgadevi¹

¹ SRM Institute of Science and Technology, Chennai, India

Abstract

The Internet of Things (IoT) has become one of the most widely used technologies in recent times. IoT devices can be enabled to collect, and exchange information in a highly efficient manner via the network. A smart object with technology and devices builds a network infrastructure that is used in a variety of areas such as mechanical, building, medical, manufacturing, entertainment, and transport. The major security issues such as confidentiality, authentication, confirmation, security systems, system configuration, data storage, and administration are the main challenges in an IoT environment. To overcome these security issues, various techniques are addressed. Initially software called an Intrusion Detection System (IDS) was used that monitors a network of malicious activity using valuable tools in IoT devices. Then, the machine technique was used to detect the attacks from the intrusion detection system to provide embedded intelligence in IoT devices and networks. Finally, Blockchain (BC) technology is gaining traction in modern IoT devices to address security and privacy challenges to provide reliable communication in an IoT environment. The aim of this work is to provide a detailed review of ML and BC techniques that can be used to develop revamped IoT security devices.

Keywords: Attack, IoT Layers, Protocols, ML Techniques.

* Corresponding author P.S. Pavithra: SRM Institute of Science and Technology, Chennai, India;

E-mail: [email protected]

1. INTRODUCTION

IoT is a network of sensors and objects that can communicate with one another without human intervention. The things in the IoT are hardware objects such as wearable sensors, that detect and collect different types of data about technology and human social activity. The Internet of Things keeps people, objects, devices, and services all interconnected at all times.

The primary objective of the Internet of Things is to create a broadband network with interrelated communication systems and applications that help physical/virtual sensors, home computers (PCs), digital phones, motorcars, and items such as fridges, washing machines, household appliances, food, and medications to be connected and embedded anywhere at all possible time and on any network. The requirements for large-scale IoT deployment are rapidly growing and eventually pose a serious security issue. Privacy, authorization, authentication, security systems, system configuration, data storage, and monitoring are the primary issues in the IoT environment [1]. IoT devices are linked to complex devices, interact with the environments, and are deployed on a wide range of unmanaged systems. They confront a number of security concerns and challenges. The Internet of Things layer is separated into four layers; its architecture is based on a standard Online communication network, and it is primarily for information transit between IoT devices. In recent years, IDS has shown to be a more reliable and efficient strategy. IDS is a technology that analyses a network for unexpected IoT device performance [2]. IDS can be set up on a single system or on multiple machines in a network. IDS provides several advantages to businesses, including the ability to detect security threats. An IDS (Fig. 1) can aid in the identification of threat types and numbers. This paper outlines a strategy for developing an IDS that employs Machine Learning (ML) approaches to detect data-based threats in order to defend against attacks in the IoT. The hostile devices carry out attacks, where data is collected in two ways: benign information during normal flow and traffic seized during threats. Machine learning techniques are built using a number of approaches to detect malicious behaviour in an IoT infrastructure. Blockchain is a distributed technology with numerous advantages, including increased security and transparency. As a result, blockchain can spearhead itself be a strong platform for payment and communication apps. Thus by using blockchain as a database to keep records of how things communicate, what state they're in, and how they connect with other IoT systems, blockchain can help to solve the majority of IoT privacy and tracing issues.

The detailed review of this paper is carried out as follows: Section 2 shows the classification of IoT layers and their protocols. Section 3 shows security issues in IoT Layers. Section 4 shows security issues occurring with the use of IDS. Section 5 discusses security issues that occur using ML techniques and Section 6 explains security issues that occur using Blockchain technology.

2. CLASSIFICATION OF IOT LAYERS

The IoT can be divided into four layers namely: Application, Middleware and Sensor Layer, as shown in Fig. (2).

Fig. (1))

IoT architecture.

Fig. (2))

IoT layers.

2.1. Application Layer

The advantages of IoT in our daily lives are prominent. Security was not a significant design when the IoT was first introduced in the late 1960s since security risks were not properly accommodated. Security has become crucial for the IoT's long-term viability and widespread adoption. IoT applications and sensors have infiltrated every part of our lives [3]. IoT has become a critical component of many healthcare contexts. IoT sensors have made their way into our living environments by paving ways to create smart home which includes Possible sources, led lights, thermostats, and other home equipment are now equipped with networking capabilities, enabling wireless remote control. Almost every household item can now be utilised in an autonomous way by operating it remotely. We are now surrounded by IoT apps and devices in our homes, automobiles, railroads, roads, transport, farming, and companies, as shown in Fig. (3).

Fig. (3))

Application layer.

2.2. Middleware Layer

Middleware for IoT devices is a technology that acts as an interface between network elements which improves the interaction of network elements that would otherwise be incompatible. Middleware combines disparate, often sophisticated, and already-existing programmes that were not linked in the first place. IoT is defined as the ability for almost anything to be interconnected and transfer data through a system. Middleware (Fig. 4) is a component of the design that enables connectivity for a large number of different things by providing a connection layer for the network layer as well as the application layers that provide solutions that enhance efficient software interactions.

Fig. (4))

Middleware layer protocols.

The Aggregator, Local Service Gateway (LSG), and IoT server are the three basic layers of IoT middleware. Sensor abstraction is provided by the aggregator which hides the hardware characteristics of the actual sensors and presents a single interface for searching and subscription to sensor data. The LSG layer receives raw data from the aggregator [4]. The LSG connects the Internet of Things system to the rest of the world. For contextual refinement and aggregation purposes, it may analyse basic data given by the aggregator. The LSG also sends the data to the IoT server, along with a data GUID, user access framework, and memory location information (human-readable names or NA). Through its edge router, programs (users) can request the IoT server about where to get information from. The data can then be retrieved from a storage place or right from the aggregator. The IoT server can select whether to perform network access internally or outsource it to the NCRS/GNRS. Huang and colleagues offer a security paradigm for the Internet of Things that aims to find the balance between security and usability. A body-area network, a home network, and a motel network are three key scenarios where customer experience is vital. A logistics IoT scenario and an office IoT situation were also studied. A survey has been carried out to fully understand consumer perceptions of the relevance of safety vs. accessibility, as well as how ready users are to sacrifice one for the other. Authentication, consistency, and accessibility were three characteristics of security that users were asked about. While different components of safety matter vary depending on the specific application, the survey results reveal that security is important to all people and in all apps. This is especially true when it comes to security and payment services.

2.3. Network Layer

The network layer controls data transfer to and from various products or applications using a range of assessment methods and techniques across wired or wireless communications [5]. The network layer takes the analysed data from the perception layer and chooses the optimal methods for transmitting it via products to IoT devices, ports, and bridges (Fig. 5).

Fig. (5))

Network layer protocols.

2.3.1. Low Power Wi-Fi

Devices that enable WiFi, like some other Wireless connections, IP communication is also supported by HaLow, which is vital for IoT systems. Let's take a look at the characteristics of the IEEE 802.11ah standard. This standard was created to deal with resource sensor network applications that demand particularly long communication. IEEE 802.11ah runs at 900 MHz in the semi-range [5]. The range is greater due to lower frequency, while greater range waves suffer from greater absorption. We can increase the range (now 1 kilometre) by lowering the frequency, but the data rate will be reduced as well, thus the compromise is not justifiable. Huge star-shaped networks, where many nodes are connected to a specific access point, are also supported by IEEE 802.11ah.

2.3.2. Zigbee

It is used for local area networks, or PANs, and is based on the IEEE 802.15.4 communication protocol standard. The Zigbee partnership, which aims to develop dependable, low-energy, and low-cost communication technologies, created Zigbee. The communication range of Zigbee devices is fairly short (10–100 metres). The Zigbee standard also specifies the details of the network and application levels. The network layer here, unlike BLE, allows for multichip routing.

2.3.3. Near Field Communication (NFC)

NFC is a very Small-range wireless transmission technology that enables portable devices to communicate with each other across a few millimetres [6]. By bringing two NFC-enabled devices close to one another, any form of data can be exchanged in seconds. RFID is the foundation of this technology. It communicates information between multiple NFC-enabled devices by utilising magnetic field fluctuations. NFC uses the 13.56 MHz frequency range, which is the same as high-frequency RFID. Active and passive modes of functioning are available. Both devices produce magnetic fields in the active mode, however, in the passive mode, only one device produces the field and the other transfers data through load modification. In rechargeable battery devices, the passive mode is useful for maximising energy efficiency. The requirements of proximity between devices have the advantage of facilitating data security such as payments. Finally, unlike RFID, NFC can be utilised for bidirectional communication. As a result, practically every smartphone on the market today supports NFC.

2.3.4. BLE

The Bluetooth Special Interest Group formed Bluetooth Low Energy, popularly called Bluetooth Smart. When compared to rival procedures, it has a shorter range and uses less energy. The BLE communication device is comparable to that used in traditional Bluetooth. It consists of two parts: the actuator and the client. The hardware and connectivity layers are implemented by the device. The computer is often a SOC with a TV. The uppermost layers' capacities are contained in the server. BLE is inconsistent with the traditional Bluetooth. There are differences between the standard WiFi and Bluetooth Low Energy (BLE).

2.3.5. Low Power Wide-Area-Networks (LPWAN)

The low-power wide area network (LPWAN) is a communication network for connecting reduced, rechargeable battery objects over long distances [7]. LPWANs that were built for M2M and Internet of Things (IoT) networks, are less expensive and consume less power than traditional wireless services. They can also connect a great variety of mobiles over a greater distance. LPWANs can accept data traffic that varies from 10 to 1,000 bytes at upload rates of up to 200 Kbps. The distance of an LPWAN can range from 2 km to 1,000 km, depending on the method. Most LPWANs have a network system, similar to Wi-Fi, in which each destination communicates to a central hub.

2.4. Sensor Layer

For all IoT systems to gather information from the environment, one or more sensors are necessary. Sensors are a crucial component of intelligent devices. Environment information is one of the most crucial parts of the Internet of Things [8], which is impossible to achieve without sensing devices. Sensors for the Internet of Things are usually compact, low-cost, and a source of energy (Fig. 6). They are limited by variables like storage capacity and easy installation. Here, a broad overview of the different types of detectors that can be used to develop effective solutions is given.

Fig. (6))

Sensor layer.

2.4.1. Mobile Phone Sensors

First, consider the omnipresent mobile phone, which contains a variety of sensors. The smartphone, in particular, is a particularly convenient and user-friendly device with a variety of built-in connectivity and information processing features. Because of the integrated sensors, researchers are expressing interest in developing smart IoT solutions employing mobile phones as a result of their growing popularity. Depending on the situation, extra sensors may be required. On the smartphone, software can be created that leverages sensor data to provide useful outcomes. The following are some of the sensors found inside a smart device [8]. A smartphone phone's sensor detects motion and acceleration. It usually measures changes in the smartphone's movement in three dimensions. Accelerometers come in a variety of shapes and sizes. An earthquake mass in housing is linked to the building using a spring-mass system in a physical accelerometer. Because the mass moves slowly and is left behind when the housing moves, the force in the spring can be linked to the acceleration.

2.4.2. Healthcare Sensors

IoT can be tremendously beneficial in health applications. Devices can be used to evaluate and monitor a wide variety of clinical works in the body. The apps can be used to monitor a patient's condition while they aren't in the healthcare setting alone. The physician, relatives, or patients can then receive real-time feedback. McGrath and Swanbill have gone to great length about the sensing devices that can be carried worn by the human body to track people's health.

2.4.3. Neural Sensors

It is now possible to study brain waves, analyse the brain's condition, and educate the mind to enhance focus and be attentive. This is known as neurofeedback. The technique used to detect mind waves is EEG or a brain parts connection. Physical interactions between nerve cells produce an electric field that can be defined in terms of wavelengths externally. Brain waves are categorised as Delta, theta, gamma, Beta, and alpha waves based on