Project Details
Projekt
RAINCOAT II: Randomization in Secure Nano-Scale Microarchitectures 2
Subject Area
Security and Dependability, Operating-, Communication- and Distributed Systems
Term
since 2020
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 440059533
The microarchitecture of a computer system is a transparent layer in modern CPUs to provide the foundation for the architectural abstraction of high-level system features and low-level hardware design. However, the abstraction is not perfect, as showcased by a massive variety of attacks exploiting microarchitectural features using timing information that is observable from software. This class of microarchitectural attacks includes prominent examples like Spectre, Meltdown, Flush+Reload, and many more. The research on microarchitectural attacks is quickly progressing, including automated approaches. At the same time, developing defenses or mitigation techniques against these side channels are very challenging. In particular, retrofitting existing microarchitectures with side-channel mitigations is often extraordinarily expensive, if at all possible. Consequently, it is vital to consider these attacks in the early design phase of future architectures, following a security-by-design approach. The transition to emerging nano-scale hardware technologies presents a pivotal point where new developments at the technology level require re-thinking of traditional design approaches and enable the development of new, efficient, and secure-by-design architectures. However, while these technologies enable the consistent performance increase that has become to be expected, they may also introduce new security risks to the microarchitecture. Our proposed research project focuses on three pillars: 1. Designing side-channel secure microarchitecture components, 2. evaluating (fault) attacks on emerging nano-scale memory technologies, and 3. enabling secure execution and maintenance on untrusted nano-scale hardware. This proposal presents a continuation of the RAINCOAT project, which is currently funded by the DFG as part of the priority programme Nano Security: From Nano-Electronics to Secure Systems (SPP 2253), with termination on October 2023. Within this application, we will present the results and findings of the first RAINCOAT project and describe open challenges for the new funding period. In RAINCOAT I, we developed building blocks for secure microarchitectures, e.g., side-channel hardened caches, TLBs, and attestation schemes that build on top of it. For the second funding period of RAINCOAT, we aim at answering the research question how we can efficiently ensure secure execution of security critical applications on modern, nano-scale microarchitectures? In addition to the continued development of secure microarchitectural building blocks, this includes the evaluation of the system as a whole and the provision of means to guarantee the protection of the most security-critical applications, such as cryptographic algorithms. For this, we will investigate the interplay of secure microarchitecture building blocks, locking mechanisms of shared resources, and runtime protected environments like TEEs.
DFG Programme
Priority Programmes
