Process mining is an established discipline for the discovery and assessment of business processes, yet traditional methods struggle to extract actionable insights from high-velocity data streams. This paper explores how process mining algorithms can leverage decentralized edge-computing architectures to provide real-time insights within high-throughput environments. We focus on the necessity of maintaining real-time viability and preventing back-pressure by processing data in close proximity to its point of origin. To address these challenges, we propose a conceptual framework for a distributed conformance checking algorithm specifically designed for the edge. We conclude by defining a research roadmap that addresses algorithmic distribution, the impact of multidimensional data stream characteristics, and self-adaptive orchestration for multi-objective optimization of accuracy, latency, and energy efficiency.
Reiter et al. (Thu,) studied this question.