Business process monitoring via decentralized complex event processing

ABSTRACT

Complex event processing (CEP) techniques have been used for business process (BP) monitoring of large organizations with high number of complex BPs and high rates of BP instance generations resulting in high number of monitoring rules and high rates of events. To circumvent the scale limitation of centralized CEP techniques, we present a decentralized CEP mechanism called dBPM that uses both task parallelism by decomposition and distribution of monitoring rules, and data parallelism by partitioning and dispatching of events. We show that dBPM is more scalable than a general distributed CEP mechanism, called VISIRI, by dynamically adapting to workload changes.