• Edge IoT devices face limited hardware, demanding efficient CEP solutions. • Study balances execution costs across CEP task graph paths via constrained programming. • Python library enables adaptive code and I/O placement for small-scale IoT. • Virtualized shared memory abstracts communication, simplifying multi-device coordination. • Optimizing critical path reduces latency and boosts throughput across devices. • Results show significant delay reduction and throughput improvement during CEP operations.
Article Summaries:
- Researchers have developed a Python library that optimizes complex event processing (CEP) on edge IoT devices by jointly deciding where to place data and code. The approach uses constrained programming to balance execution costs across different paths of a CEP task graph, targeting critical‑path performance. By abstracting communication details and virtualizing shared memory between devices, the library enables small‑scale IoT nodes to adaptively reallocate I/O assignments. Experiments show that this combined optimization reduces latency and increases throughput during CEP operations across multiple devices, demonstrating a practical method for improving edge‑computing efficiency.
Sources: