Complex event processing: Current status and considerations for edge deployment
Styliani Kyrama, Anastasios Gounaris · Future Generation Computer Systems · 2025
• Main differences between CEP and SP systems with a focus on edge computing • Overview of optimizations, disorder and failure handling in CEP systems • Considerations and challenges regarding the deployment of a CEP system on the edge The need for real-time data processing and analysis has led to the emergence of two important paradigms; generic Stream Processing (SP) and Complex Event Processing (CEP). While SP focuses on analyzing continuous data streams in real-time to extract insights, CEP specializes in identifying meaningful event patterns and correlations. This work highlights the distinctions between CEP and SP, including the types of data processed and the operators used, as well as the emphasis on pattern detection. It also provides an overview of the existing optimization techniques adopted by SP and CEP systems, and discusses critical requirements for CEP systems to ensure result correctness, such as handling inconsistencies, out-of-order and duplicate data, effective state management, and security considerations. Furthermore, our work investigates the deployment of CEP systems in edge settings, to address latency, security, and privacy concerns, highlighting the challenges specific to these systems. We conclude with a vision for edge-based CEP engines, identifying research gaps, and offering guidance for further exploration.