Attacks in the IoT
We can broadly classify security attacks in the IoT with the help of the following four domains:
When an attacker maintains a physical closeness with the network or other devices of the system, then physical attacks are launched. Physical attacks include:
When a device (e.g., RFID) or any transmission link is enhanced physically, then tampering is introduced . Confidential and sensitive information can be acquired as the result of tampering. Research has been done to identify the vulnerabilities in popular physical devices (such as smart meters, IP cameras, and Amazon Echo). A camera’s password can be acquired by an attacker, regardless of its length and a configuration that is confidential and sensitive to a particular user. A physically unclonable function (PUF) is the proposed countermeasure for this attack. The idea of a PUF was proposed for small-sized IoT devices, in order to exploit the integral instability of integrated circuits (IC). A PUF uses a challenge-response mechanism, where the physical structure at the micro level of the device primarily decides the output of the system. Thus, the PUF is an efficient measure to eliminate attacks such as tampering.
In order to hinder a communication, instead of sending radio frequency (RF) signals, an attacker creates and transmits noise signals to launch DoS attacks on RFID tags. This is known as RF interfacing/jamming. Hindering or jamming communication is the prominent effect of this attack. A customizable and trustable end device mote (CUTE mote) is the proposed countermeasure used for this attack. With the help of a CUTE mote, in order to achieve productivity and overall performance, a solution of the device is presented in the work . With a hardcore microcontroller unit (MCU) and an IEEE 802.15.4 radio transceiver, we have some essential elements of the architecture, i.e., a reconfigurable computing unit (RCU).
Fake Node Injection
In order to control data flow, a fake node is dropped by an attacker into the connecting legal nodes of the entire web, and this is known as a fake node injection. Managing the flow of data is the effect of this attack. An attacker can acquire control of the process of any data. Several physical devices are vulnerable to this attack. Pathkey is the proposed countermeasure against this attack. In today’s world, distributed IoT applications are a fundamental component of great importance. The establishment of secure links between each and every sensor node and end user is extremely important, in such an environment.
Those attacks in which battery-powered devices are continuously used, by providing them with incorrect inputs by an attacker, are known as sleep denial attacks. Sleep denial attacks are vulnerable to node shutdown, which is their main effect. A CUTE mote and support vector machine (SVM) are the proposed countermeasures for this attack. A CUTE mote is beneficial to prevent sleep denial attacks, because of its heterogeneous structure. An SVM has been designed that uses medical access patterns from a patient’s device. Because of its classification algorithm, an SVM is capable of determining resource exhaustion, which makes it invulnerable to sleep denial attacks.
Those attacks that damage an entire system by orchestrating the IoT network are termed network attacks. Without even being close to the network, these attacks can be launched easily. Common network attacks include:
Traffic Analysis Attack
Attackers acquire confidential information even without being close to the network, so that they can obtain information about the network. These attacks are vulnerable to data leakage, i.e., unauthorized access to network information. The efficient and privacy preserving traffic obfuscation framework (EPIC) is the proposed countermeasure for traffic analysis attacks. Jianqing Liu et al.  have proposed the EPIC framework, so that smart homes could be protected from traffic analysis. It assures the pseudonymity of the flow of traffic to a specific smart home and also between source and destination. This framework acts as a secure multihop routing protocol, assuring strong privacy prevention.
In RFID spoofing, in order to acquire the data stamped on the RFID tag, an RFID signal is first spoofed by the attacker. The attacker then sends its data by posting it as valid, with the help of the original tag ID. This attack is designed to manage and modify data (i.e., reading, writing, and deleting). The physically unclonable function (PUF) based on SRAM is the countermeasure against RFID spoofing. A PUF based on an on-board SRAM has been invented, which fabricates a unique device ID with the help of the unique device footprint. The possibility of the impersonation of an ID by an adversary can be minimized by using device ID matching, which prevents the risk of spoofing, as well as fraudulent access.
Routing Information Attacks
Direct attacks where the attacker creates a nuisance through activities such as creating routing loops, sending error messages and spoofs, or altering routing information, are known as routing information attacks. Routing loops are the prominent targets of these attacks. Hash chain authentication is the proposed countermeasure for routing information attacks. Hash chain authentication is required in order to efficiently deal with routing attacks, by preventing malicious codes from exploiting control messages. Selective forwarding and sinkhole attacks can be effectively reduced by using hash chain authentication and rank threshold in combination.
In a man-in-the-middle (MitM) attack, an attacker can gain access to the private data of any user by eavesdropping or monitoring the communication between two IoT devices. The violation of the data privacy of any user is the prominent effect of MitM attacks. An IoT system can be seriously impacted by MitM attacks. For example, an attacker can take control of a smart actuator in an industrial IoT setting . They can potentially damage an assembly line, by knocking an industrial robot out of its designed lane and speed limit. MQTT and inter-device authentication are the proposed countermeasures for this attack. MQTT ensures device-to-device (D2D) communication by using key-police (KP) to implement elliptic curve cryptography (ECC), which prevents MitM attacks.