| A Wireless Sensor Network (WSN) is composed of a large number of sensor nodes. These nodes have the capability of sensing, computing and communicating. In the year 1999,"Business Week" of America considers WSNs as one of the 21 most important ideas for the 21 century. WSNs become very popular in recent years and are broadly used in many areas, including forest fire detection, military surveillance, environmental monitoring, etc. With the development of WSNs, great progresses have been made in the area of sensornet OS research, e.g., TinyOS, SOS, Contiki, Mantis and RETOS.Networking support plays a central role in the design of a sensornet OS. A good design and implementation of a networking stack will improve communication performance, reduce energy consumption, and simplify user application development. However, it is challenging to implement a flexible, energy efficient, easy-to-use networking stack on sensor nodes with severe resource constraints.In this thesis, we analyze networking stacks of the state-of-art sensornet OSes, and then present the main principles of our layered design networking stack for the SenSpire Operating System. SenSpire OS's networking stack consists of three layers, i.e., the radio layer, the resource layer, and sensornet layer. The radio layer implements device specific functionalities and provides unified interface to the resource layer. The resource layer sits between the radio layer and the sensornet layer. It uses the interface provided by the radio layer, provides resource management and concurrency control for the upper layer. The network support layer provides useful services, e.g., neighborhood management and buffer management, to implement specific network protocols.At the end of this thesis, we show experimentation results of SenSpire OS's networking stack, including transmission throughput, energy consumption, flexibility, etc. We demonstrate that our networking stack is easy-to-use, and is promising in simplifying WSN application development. |
PDF Full Text Request