Visible to the public Semantics-Directed Prototyping of Hardware Runtime Monitors

TitleSemantics-Directed Prototyping of Hardware Runtime Monitors
Publication TypeConference Paper
Year of Publication2018
AuthorsHarrison, William L., Allwein, Gerard
Conference Name2018 International Symposium on Rapid System Prototyping (RSP)
Date Publishedoct
ISBN Number978-1-5386-7557-1
KeywordsBiomedical monitoring, building memory protection mechanisms, composability, embedded hardware, embedded processor, Embedded systems, formal verification, Hardware, hardware monitor, hardware runtime monitors, hardware security, hardware-based protection, high level synthesis, Human Behavior, human factors, Instruction sets, Model driven development, Monitoring, multiple memory accesses, pubcrawl, rapid-prototyping runtime monitors, reconfigurable architectures, resilience, Resiliency, rop attacks, ROP-style code reuse attacks, Runtime, Scalability, security of data, Semantics, semantics-directed prototyping, software monitor, software-based attacks

Building memory protection mechanisms into embedded hardware is attractive because it has the potential to neutralize a host of software-based attacks with relatively small performance overhead. A hardware monitor, being at the lowest level of the system stack, is more difficult to bypass than a software monitor and hardware-based protections are also potentially more fine-grained than is possible in software: an individual instruction executing on a processor may entail multiple memory accesses, all of which may be tracked in hardware. Finally, hardware-based protection can be performed without the necessity of altering application binaries. This article presents a proof-of-concept codesign of a small embedded processor with a hardware monitor protecting against ROP-style code reuse attacks. While the case study is small, it indicates, we argue, an approach to rapid-prototyping runtime monitors in hardware that is quick, flexible, and extensible as well as being amenable to formal verification.

Citation Keyharrison_semantics-directed_2018