Research And Implementation Of Device Management Solution Based On Internet Of Things Service Platform

With the maturity of sensor and wireless communication technology, IoT (Internet of Things) application has been extensively spread, which brings people great convenience. However, with more and more diverse devices appearing in people’s lives, the device management is faced with great challenge. In many real scenarios, there are a mass of devices deployed, among which existing a considerable number of capability-constrained ones, whose underlying protocols and collected data are very distinct from each other. Under this circumstance, the traditional device management in IoT cannot fully meet the requirement any longer. Therefore, this thesis designs lightweight device management architecture for the resource-limited scenario and implements a uniform management platform fit for capability-constrained devices.The research work of this thesis contains three parts. Firstly, we design lightweight device management architecture by making reference to the OMA LWM2M standard. The proposed architecture owns three layers:DM Client layer, Intermediate layer and Web application layer. The DM Client layer is responsible for the abstract description of resources. The Web application layer is the interface between users and the system. The Intermediate layer acts as a bridge to implement seamless connection of web application and IoT system.Secondly, we design and implement each layer and their modules in details. On the DM Client layer, we build the Object-Resource model and design the URI to describe resources. On the Intermediate layer, we design and implement the functional interfaces, then put emphasis on the research of CoAP in the protocol stack module and propose an improved reliability control algorithm implementation. Based on the open-source project jCoAP, we implement the transformation of HTTP and CoAP, so that the web client based on HTTP can access the device network based on CoAP transparently. On the Web application layer, we use the lightweight frame Struts, Spring and Spring-Data-MongoDB to implement business logic processing and MongoDB accessing. In addition, data visualization is implemented by D3.In the end, we implement a prototype system based on the proposed architecture and conduct some experiments to verify the functional and performance feature by using the wireless sensor network nodes. The result shows that our proposed device management system fulfilments the uniform management function and also improves the system performance by reducing the response time in resource-constrained scenario.