In conventional wireless sensor networks, sinks collect data from sensors and/or aggregators and interact with end-user applications via centralized gateways. However, there exist application scenarios where this modus operandi may not be suitable. In such cases it is preferable that the applications interact directly with the sensors. This paper proposes a Web- service based architecture with an overlaid UDDI for these cases. It also presents simulation results for these scenarios to demonstrate performance gains.

It is foreseen that several wireless Sensor Networks (WSNs) will be deployed in the very near future to provide various sensing services to end user applications. A wide range of novel end-user applications will emerge. In some application scenarios, the sinkless model is appropriate. In this model, the applications interact directly with the sensor nodes, without going through a gateway. However, the current sensor nodes use proprietary protocols and data access frameworks. This propriety will hinder the speed of evolution of applications. There is a need for a framework that is well known to application developers and is based on standard protocols. This paper presents an embedded Web Services platform residing on the sensor nodes, allowing direct interaction between applications and individual sensor nodes. The lessons learned indicate that web service platform on sensor node is feasible, despite the nodes' limited resources and the complexity of the framework.

Applications are the ultimate consumers of the information collected by sensors. There exist several frameworks for the interactions between end-user applications and sensors. They range from low-level APIs to databases and include Web Services. This paper is devoted to the use of Web Services for bridging end-user applications and wireless sensors networks. Its first contribution is a systematic evaluation of the current frameworks for the interactions between end-user applications and wireless sensors networks. The evaluation shows the potential of Web Services, compared to the other frameworks and motivates their usage in the case study. The second contribution is the definition of Web Services for the I-centric telecommunication services, and their implementation in a wireless sensor network that does not support Web Services. We demonstrate that Web Services are very promising as "bridges" and share the lessons we have learned.

It is foreseen that in the near future, several Wireless Sensor Networks (WSN) will be deployed to provide data and services to end user applications. WSN will play an important role in context aware applications and services, which adapt to situations of the surrounding environment and that of their host devices. However, conventional WSNs are built using proprietary protocols and use proprietary data access frameworks. They necessitate the presence of gateways or protocol converters between the WSN and application devices, and require special expertise from application developers. Further, dealing with heterogeneous sensor networks is difficult, as the protocols and data access mechanisms are different. There is a need to have a framework that is based on open standards to simplify application development and deal seamlessly with heterogeneous WSNs. This thesis proposes a Web Services based sinkless architecture for WSN as a potential solution. A prototype of an embedded web services platform is also built. It allows application devices to interact directly with the sensor nodes without going through a gateway. The thesis includes a survey and evaluation of several frameworks to highlight the potentials of web services. The use of Web Services will also attract a larger pool of application developers, leading to more innovative applications. Further, as the framework is built over IP, WSNs will no longer be treated as foreign islands. They will be part and parcel of ICT infrastructure in enterprises, as well as fit in easily in Value Added Services infrastructures for Next Generation Telecommunication Networks.