| In recent years, the rapid development of Wireless Sensor Network (WSN) and Cloud Computing turns out to be a new character of the internet of things. People want to enjoy WSN resources freely without any limitations on time, location or channel. Generally, WSNs have comparatively lower data processing ability, lower bandwidth capacity and lower transmission ability than other networks. The special many-to-one communication and multi-hop transmission operation mode makes WSNs more prone to congestion. Control mechanism of congestion is the key technique affecting the energy utilization and the network QoS. With the increase of users of WSN, the performance of the server-side data platform is limited. To boost the overall performance, not only the hardware of server port but also the software should be improved. The purpose of this thesis is to improve the performance of the server and control network congestion by designing the software properly.The main contributions of the thesis are summarized as follows:1. We analyze the existing congestion control mechanism of WSNs and study the causes of congestion. Then, we propose to control the congestion in WSNs by using data platform buffer module.2. The influence of threading technology and queue technology on the data processing is analyzed in depth, which is the theoretical basis of our proposed scheme for WSN congestion control. Based on the data management theory of the data platform, we design a new control structure, as an alternative of the original, using the thread pool and queue pool technologies..3. Having analyzed the queue pool and thread pool technologies, we employ the concepts of thread pool, message queue, and priority scheduling in our design of the buffer module of the data platform. All functional modules and interfaces have been designed in detail and implemented on selected software platform. Simulation results show that the designed data platform buffer module can improve the performance of WSN transmission effectively. |
PDF Full Text Request