Resumen de Useme: a service-oriented framework to facilitate wsan application development
WSANs are a new and promising technology but in order to fully develop their possibilities, higher levels of abstraction are needed. SOAs (Service Oriented Architectures) have proven to be an efficient and inter-operable solution to communicate remote nodes. This thesis has a framework based on services which allows the programmers to create applications for WSANs without having to worry about low level and repetitive tasks present in all WSAN applications. The framework has interesting features that make it suitable for a wide range of applications. USEME is platform independent, a combination of macro-programming and node-centric programming and it has group hierarchy architecture with the possibility of defining group constraints. It also provides programmers with a high level abstract language for service definition and composition, real-time support that allows programmers to define QoS constraints in the communication between services and a visual tool that assists programmers to specify the global behavior of the network. Another interesting point of the framework is the novel mechanism to wirelessly reprogram the nodes of a sensor network by using neural networks. This reprogramming process is not only regardless of the operating system executed in the sensor nodes but also, in contrast to many traditional reprogramming proposals, small code changes can be carried out without having to reprogram the whole application.
It is worth remarking that one of the most important parts of any framework is the middleware layer used to make their offered facilities possible and transparent to the users. USEME framework uses a middleware which integrates a novel communication hierarchical protocol called HERO that, thanks to its design and features is a fault-tolerance, energy-efficient and reliable protocol in the context of wireless sensor and actor networks. HERO organizes the nodes of the sensor networks in groups through two phases: discovery phase and joining phase. During the first phase, sensor nodes discover which are the shortest paths to reach their group leaders. Once, the sensor nodes know how to reach its group leader, they ask it for permission to belong to its group through the joining phase. Group leaders take advantage of this phase to discover which are the clue-nodes between them and their sensor nodes. They will allow them to estimate where their sensor nodes are located. Experiments have shown how clue-nodes help to save and distribute the energy consumption throughout the network. Regarding the packet delivery mechanisms, HERO not only allows developers to send data both ways (from sensor nodes to its cluster-head and viceversa) but it also allows them to specify the desired reliability level in a quantitative way. Furthermore, we also introduce the concept of ¿memory path¿ which allows nodes to save more energy when nodes need an ACK packet to confirm that the information has arrived correctly at its destination. In addition, HERO allows users to use a novelty reliable and energy-efficient mechanism transmission. It permits developers to numerically (0 to 100%) set the desired reliability level between two nodes which are N hops away from each other. Basically, the algorithm is able to know and achieve the needed reliability of the intermediate nodes used to send information between nodes (far away from each other) with a reliability previously specified by the user.
Finally, it is well known that more and more the phenomenon ¿Internet of the Things¿ is increasingly having an impact in our day to day lives. Because of this, in the near future practically of the ¿things¿ (cars, mobiles, clothes, light streets, buildings, . . . ) will be ¿talking¿ with each other through the Internet and we are convinced that our service model is ideal for use in these promising new networks.
