Migrating Complex Event Processing in the Web of Things

Through the IoT and Internet applications integration emerges the Web Of Things (WoT). Currently, WoT solutions generate data that can be used for decision making. Among the techniques of data flow analysis, one that has gained prominence is Complex Event Processing (CEP). Considering a WoT device may have connectivity, memory and processing power enough to perform CEP locally, it is possible to reduce the response time of data processing as a whole. Based on this premise, this work investigates the execution of a CEP system in an WoT device, analyzing the CPU and memory consumption in three scenarios: Mist (in the device itself), offloading in the Fog and offloading in the Cloud through WebSocket and HTTP protocols. The results show we can run CEP code on a WoT device, but its performance depends on the time window processing and the throughput of the data that reaches the device.