物聯網中介軟體中的複雜事件處理系統

Abstract

近年來，關於物聯網的研究與日俱增、相關應⽤設備與系統也漸漸出現在市⾯上，然⽽，現今物聯網應⽤系統發展中⾯臨了幾個問題：缺乏複雜事件之概念，無法有效分析由感測設備來的資料成為人類所能理解的事件。為解決這些問題，在此論文中我們將提出一個複雜事件處理系統，處理從感測設備收集到的資料，將其轉換為事件，再經由複雜事件處理引擎利用事件模式轉換為複雜事件，此外使⽤者也能⾃行定義事件模式並應用於不同情境。本論文實作複雜事件處理引擎：使⽤者可定義複雜事件模式，同時複雜事件處理系統將gateway收集之感測器資料轉換為事件輸⼊處理引擎，引擎則採⽤樹結構⽅法偵測出複雜事件。最後對複雜事件處理系統進行情境驗證與效能評估。

In recent years, Internet of Things has been attracting enormous attention not only from the research community but also from the industry sectors. However, the development of the software for Internet of Things has faced many problems, in particular, lacking the concept of complex event processing makes it hard to analyze data derived from the transducers. To solve this problem, we propose a complex event processing system which can read data from sensors, transform them to atomic events, put events into complex events processing engine, and output the complex events predefined. We developed the complex event process engine: user can define the complex event pattern, transducer data collected from the gateway will be transformed to events and last the tree-based complex event processing engine will find out the complex events. Finally, we present the scenario verifications and performance evaluations.