Optimising Security, Power Consumption and Performance of Embedded Systems

Increased interest in multicore systems has led to significant advancements in computing power, but it has also introduced new security risks due to covert channel communication. These covert channels enable the unauthorized leakage of sensitive information, posing a grave threat to system security. Traditional examples of covert channel attacks involve exploiting subtle variations such as temperature changes and timing differences to clandestinely transmit data through thermal and timing channels, respectively. These methods are particularly alarming because they demand minimal resources for implementation, thus presenting a formidable challenge to system security. Therefore, understanding the different classes of covert channel attacks and their characteristics is imperative for devising effective countermeasures. This thesis proposes two novel countermeasures to mitigate Thermal Covert Channel (TCC) attacks, which are among the most prevalent threats. In the first approach, we introduce the Selective Noise-Based Countermeasure. This novel technique disrupts covert communication by strategically adding a selective noise (extra thread) to the temperature signal to generate more heat and change its pattern. This intervention significantly increases the Bit Error Rate (BER) to 94%, thereby impeding data transmission effectively. Building upon this, the second strategy, termed Fan Speed Control Countermeasure, dynamically adjusts fan speed to reduce system temperature further, consequently decreasing the thermal signal frequency and shutting down any meaningful transmission. This methodology achieves a high BER (98%), thereby enhancing system security. Furthermore, the thesis introduces a new threat scenario termed Multi-Covert Channel Attacks, which demands advanced detection and mitigation techniques. To confront this emerging threat, we propose a comprehensive two-step approach that emphasizes both detection and tailored countermeasures. This approach leverages two distinct methodologies for implementation, with the primary goal of achieving optimal performance characterized by high BER and low power consumption. In the first method, referred to as the double multi-covert channel, we employ two distinct frequency ranges for the timing and thermal covert channels. Through extensive experimentation, we demonstrate that this approach yields a high BER, providing a formidable challenge to various defense strategies. However, it is noteworthy that this method may potentially lead to overheating issues due to the increased operational load. Alternatively, our second method, the single multi-covert channel, employs a single frequency range for data transmission. Notably, this approach addresses the overheating concerns associated with the double multi-covert channel, thereby reducing power consumption and minimizing the risk of system overheating. The experimental results presented in this thesis demonstrate the efficacy of the proposed strategies. By adopting a two-different approach, we not only enhance detection capabilities but also mitigate potential risks such as overheating. Our findings contribute significantly to the ongoing discourse on covert channel attacks and offer valuable insights for developing robust defense mechanisms against evolving threats. By providing insights into both traditional and emerging covert channel threats in multicore systems, this thesis significantly contributes to the field of multi-embedded system security. The proposed countermeasures demonstrate tangible security improvements, while the exploration of multi-covert channel attacks sets the stage for detection and defense strategies.