Security Assurance for System-on-Chip Designs With Untrusted IPs

Abhishek Basak,Swarup Bhunia,Thomas Tkacik,Sandip Ray
DOI: https://doi.org/10.1109/tifs.2017.2658544
IF: 7.231
2017-07-01
IEEE Transactions on Information Forensics and Security
Abstract:Modern system-on-chip (SoC) designs involve integration of a large number of intellectual property (IP) blocks, many of which are acquired from untrusted third-party vendors. An IP containing a security vulnerability-whether inadvertent or malicious-may compromise the trustworthiness of the entire SoC, e.g., by leaking sensitive information or causing execution failures at key points. Existing functional validation approaches, post-manufacturing tests, and IP trust verification techniques are inadequate to accomplish comprehensive system-level security assurance in the presence of untrusted IPs. In this paper, we analyze security issues at the SoC level caused by untrusted IPs. We also propose a novel, resilient SoC security architecture to ensure trusted SoC operation with untrusted IPs. Our architecture realizes fine-grained IP-trust aware security policies in an efficient security policy checker that enables run-time monitoring of security issues arising from untrusted IPs. It also exploits on-chip design-for-debug architecture to ensure trusted information flow from IP blocks to the security policy checker. Unlike existing solutions to the untrusted IP problem, which rely on verification of IP trust before they are integrated into an SoC, the proposed approach follows a fundamentally different architecture-level solution based on run-time resilience. We demonstrate the effectiveness of this framework for system protection using several illustrative practical use cases. We also provide experimental results to show that the overhead of the proposed architecture is modest on representative SoC designs.
computer science, theory & methods,engineering, electrical & electronic
What problem does this paper attempt to address?