物聯網服務協作代表將不同物聯網裝置所提供之不同的功能整合,以達到某個特定的目的。由於雲端提供豐沛之運算與儲存資源,過往大部分的物聯網服務協作會架至於雲端。然而,雲端與裝置間往往存在不可避免之訊息延遲。而某些物聯網應用可能無法忍受這樣的延遲。因此,如何將物聯網服務協作從雲端遷移至更靠近物聯網裝置的霧端變得格外重要。在本論文中,我們提出了一種基於模式驅動工程之霧端物聯網協作方法。本論文採用物件導向之手法對複雜環境進行塑模。此模型主要包含四大部分:虛擬物件、物聯網服務代理人、霧端/雲端服務代理人與服務協作引擎。最後,我們設計了一個物料倉儲的案例以說明此模型之可行性。
IoT service orchestration, which means integrating different functionalities provided by IoT devices to realize a specific task, is a prevalent issue. In the past, the work of service orchestration is usually done on cloud because there are abundant resources provided on cloud. However, the latency between the edge devices and the cloud is inevitable and some latency sensitive applications may not be able to tolerate it. Therefore, the concept of fog computing is adopted and how to orchestrate the IoT services on fog becomes more and more important. In this thesis, a model driven approach of IoT service orchestration on fog is proposed. The concept of objected-oriented principles is adopted to model the complicated environment. The model includes four parts, which are virtual objects, IoT service agents, fog/cloud service agents and the service orchestration engine. In addition, a use case of material storage is designed in order to explain the usage of this service orchestration model.