Understanding IoT Device Security Challenges and Solutions

• IoT devices have limited computing resources, and complete operating systems are usually not deployed on them, so IoT devices rarely use binary security hardening measures commonly found on PCs and smartphones, such as Address Space Layout Randomization (ASLR), non-overlapping writable and executable memory, and Stack Smashing Protection (SSP). This makes software vulnerabilities on these devices more easily exploitable by attackers to gain control over the devices.

• Similarly, IoT devices typically do not have complex security mechanisms such as firewalls, and their security heavily relies on the safety of the surrounding network environment.

• Some (especially low-end) IoT devices have overly simplistic security mechanism designs, and the code on these devices has not undergone thorough security audits, making them more susceptible to security weaknesses, increasing the risk of security attacks.

• Due to the omnipresence of IoT devices and their long online durations, they are more easily scanned and attacked. Especially for devices that are long exposed to the public network, such as routers and Raspberry Pis, these devices are prime targets for hackers to research, scan, and attack. Once an IoT device is compromised and enters a botnet, it can cause long-term and persistent damage to other devices on the same network, and it may be difficult for the device owner to detect.

The characteristics of IoT devices make them more sensitive to physical layer side-channel attacks. For example, the power consumption of IoT devices is strongly correlated with the CPU’s execution state, making it easy for attackers to infer the CPU’s execution state by measuring changes in the device’s power consumption, potentially leaking confidential information such as user passwords. Additionally, IoT devices are also more susceptible to Denial of Service (DoS) attacks; for instance, an attacker can repeatedly wake an IoT device to reduce its lifespan.

Security at the perception layer. The main threats faced by the perception layer include the following types:

• Physical Attacks. Once an attacker has physical access to a device, they can directly extract sensitive data from the device or modify the device’s program to insert malicious code. Such attacks require physical contact with the device, making them relatively easy to prevent; however, device manufacturers also need to consider the physical security of the device itself, such as avoiding leaving debugging interfaces on the device’s circuit board.

• Identity Spoofing. In IoT networks lacking effective authentication measures, identity impersonation is likely to occur, where malicious devices can steal sensitive data through this method. Additionally, insecure device initialization and pairing processes are also significant sources of security threats; for example, an attacker can inject malicious programs or configurations into the router during its first boot, when no password is set. To prevent such attacks, device manufacturers should carefully consider potential security issues in the authentication process and implement effective authentication mechanisms.

• Denial of Service (DoS) Attacks. Attackers can exhaust a device’s computing power by keeping it in a high-power state for an extended period. Completely preventing such attacks is very difficult, but device manufacturers can minimize the damage caused by such attacks through reasonable security design. For example, for operations that consume a lot of computing power, effective client authentication mechanisms should be implemented, or the frequency of client operations should be limited, etc.

• Data Transmission Attacks. For instance, attackers can intercept the data transmitted by devices or conduct Man-in-the-Middle (MitM) attacks. General preventive measures against such attacks include increasing encryption and security verification mechanisms during data transmission, such as WPA (Wi-Fi Protected Access) in Wi-Fi, and EEA (EPS Encryption Algorithm) in LTE (Long Term Evolution).

• Data Leakage: Design flaws in programs may lead to the leakage of users’ private data to third parties. To avoid such vulnerabilities as much as possible, thorough security audits and evaluations of the program code on the devices are necessary.

• Denial of Service Attacks: Similar to the denial of service attacks in perception layer security, and will not be repeated.

• Remote Code Execution: By exploiting security vulnerabilities in device programs, attackers can control devices to perform unexpected actions or even execute arbitrary code uploaded by the attacker. Preventive measures against such security threats are similar to those for traditional vulnerabilities, mainly including: strengthening code security audits, conducting fuzz testing, and adopting binary security hardening measures such as ASLR. However, when preventing such vulnerabilities, the resource limitations of the devices themselves must also be considered.