Area-efficient event stream ordering for runtime observability of embedded systems

摘要

The complexity of embedded systems presents key challenges for in-situ monitoring and analysis of complex hardware and software interactions. System-level observation methods have enabled nonintrusive runtime methods for monitoring this complex behavior across hardware and software boundaries and for deeply embedded components. Previous system-level observation methods utilized an efficient pipelined hardware architecture to ensure events are reported in-order based on the event occurrence. While providing high throughput for reporting events, this approach requires significant area resources. In this paper, we present an area-efficient event stream ordering technique that significantly reduces area requirements with tradeoff in event stream throughput.