EPA-RIMM, or "Extensible Performance Aware Runtime Integrity Measurement Mechanism", is "A Framework for Dynamic SMM-based Runtime Integrity Measurement" for x86 CPUs or, in layman's terms, checking a system/hypervisor during runtime for any unexpected changes in order to detect persistent threats (like rootkits) early. It relies on SSM (System Management Mode), a feature of x86 CPUs that suspends all normal execution including the OS.
Previous work utilizing SMM in runtime integrity checking had high performance costs to the system, EPA-RIMM's primary innovation is dividing long-running checks into much smaller tasks.
I worked on EPA-RIMM as a Master's project for one term mostly centered around writing a test suite and fixing any bugs those tests found. I also handled converting Python sections of the codebase from Python 2 to 3, which was mostly automatic with 2to3 but required some manual effort. My work was mostly confined to the Python "communication layers", but debugging processes involved poking around in the C code from time to time.
Papers
EPA-RIMM has been the subject of a couple published papers, all dating before my contribution, but also all still relevant to the goals of the project today.
EPA-RIMM: A Framework for Dynamic SMM-based Runtime Integrity Measurement (2018) EPA-RIMM : An Efficient, Performance-Aware Runtime Integrity Measurement Mechanism for Modern Server Platforms (2019)