UC San Diego

Wireless sensor networks have become quite pervasive in the last few years. As their technology has matured, they have transformed from an academic research area to a viable means of solving practical engineering problems. This dynamic field has several vendors who can provide the necessary software and hardware infrastructure in order to get up and running quickly. While it is relatively easy to get up and running with a 'laboratory' setup, it is a completely different story when it comes to deploying a real world wireless sensor network. Any real world deployment, whether it is outdoors or indoors, has its own unique challenges as the scale of the deployment increases. We have observed these challenges, especially involving indoor deployments while working on several research projects in our lab. Our previous projects required a large and distributed deployment of an indoor wireless sensor network. Those ad hoc deployments were done manually, making us realize the need for management primitives to do things more efficiently. The solution as proposed in this thesis, is a set of technology agnostic management primitives that help in overall management of wireless sensor networks. We used ZigBee as it is quite popular for smart building applications. Leveraging our management primitives we built features to address these challenges of administering a wireless sensor network in a building. We have used these primitives to deploy a small scale sensor network comprising of 100 nodes, which we then used to demonstrate their benefits as well as evaluate our deployment