Computer chips are splitting into smaller, specialised tiles that plug together like Lego bricks — and the security tools designed for single, monolithic chips cannot protect them. This matters because the chip industry is rapidly adopting "chiplet" designs, where multiple smaller chips from different manufacturers are combined on a single package. Standards like Universal Chiplet Interconnect Express (UCIe) make this interoperability possible, but they also create new attack surfaces. Malware can hop between chiplets, hide in one component while attacking another, or evade traditional hardware-based detectors that only monitor the CPU. Current security methods leave most of the system exposed. Chiplet-Guard aims to build a security framework designed from the ground up for these modular architectures. It will identify new types of chiplet-specific threats, develop an AI-driven monitoring system that watches all chiplets simultaneously, and create hardware mechanisms for secure communication between them. If successful, the project could protect critical infrastructure, industrial IoT systems, and connected medical devices that increasingly rely on chiplet-based computing — systems most people never think about, but which quietly keep power grids, hospitals, and factories running.
View original technical description
Chiplet-Guard is a pioneering research initiative addressing the security challenges of chiplet-based computing architectures. As the demand for high performance and energy efficiency grows, traditional monolithic System-on-Chip (SoC) designs face physical and economic limitations. Chiplets, which consist of multiple dies integrated on a single substrate, offer a scalable solution by allowing different process technologies to be used for different components. This new architecture is being rapidly adopted by the industry, with standards like Universal Chiplet Interconnect Express (UCIe) enabling interoperability among various vendors. However, the modular and heterogeneous nature of chiplet-based systems introduces new security vulnerabilities and attack surfaces, necessitating a proactive approach to security. The Challenge: Current security paradigms and malware detection methods, primarily designed for monolithic chips, are inadequate for chiplet-based systems. These systems are vulnerable to complex attacks that exploit their modular nature, such as malware offloading and cross-component threats that evade traditional detection mechanisms. Existing hardware-based malware detectors (HMDs) typically monitor a single component, like the CPU, leaving other parts of the system exposed. There is a critical need for a holistic security framework capable of detecting and responding to threats across the entire chiplet ecosystem. Aims and Objectives: Chiplet-Guard aims to develop a comprehensive security framework specifically tailored for chiplet-based architectures. The project has four main objectives: New Malware Threats in Chiplets: Identify and characterize the new types of malware threats unique to chiplet systems, focusing on how these threats exploit the heterogeneity and complexity of these architectures. AI-based Global System Security Monitoring (AI-GSM): Develop a global, robust AI-driven security monitoring system that integrates data from all chiplets to detect and mitigate security threats effectively. Architectural Support for Efficient Security: Create hardware-aware security mechanisms that facilitate secure communication and mutual trust among chiplets, ensuring efficient and interoperable threat detection. Demonstrator: Build a proof-of-concept system to showcase the proposed security solutions, validate the effectiveness of the Chiplet-Guard framework, and demonstrate real-world applicability. Potential Applications and Benefits: The Chiplet-Guard project promises significant advancements in cybersecurity for next-generation computing platforms. By addressing the specific security needs of chiplet-based architectures, the project will contribute to safer and more resilient computing systems across various sectors, including industrial IoT, connected health, and critical infrastructure. The development of a global security monitoring approach will enhance threat detection capabilities and provide a template for future security frameworks in modular computing systems. The outcomes of this research will not only protect sensitive data and systems but also pave the way for new standards and policies in hardware security, influencing both industry practices and regulatory frameworks. Conclusion: Chiplet-Guard's research is timely and essential, aligning with global cybersecurity priorities and addressing a critical gap in the security of modern computing architectures. The project’s success will position the UK as a leader in chiplet technology innovation, with widespread benefits for academia, industry, and public policy. The initiative will ensure that chiplet-based systems are secure by design, safeguarding future generations of computing platforms from evolving cyber threats.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know