Task allocation in the Internet of Things

The last few years have been involved by the technological revolution represented by the Internet of Things (IoT). The IoT vision aims to interconnect devices with different capabilities such as sensors, actuators, Radio Frequency Identification (RFID) tags, smart objects (e.g. smartphones), and servers, within the same heterogeneous network. The aim is to enable the network objects to dynamically cooperate and make their resources available, in order to reach a goal, i.e. the execution of one or more applications assigned to the network. As known since its invention, the Internet interconnects nodes with dissimilar characteristics without central authorities by introducing some simple yet effective protocols that allow for nodes' interoperability so that information is successfully exchanged and services are provided by servers to clients and among peers. Fortunately, the same happens among objects in the IoT so that interoperability is assured and the data sensed by objects distributed and connected to the physical world is now available for the benefit of the human users. The realization of the IoT paradigm relies on the implementation of systems of cooperative intelligent objects with key interoperability capabilities. However, to reach this goal, it's important to consider some key features that characterize many IoT objects: i) available nodes' resources (electrical energy, memory, processing, node capability to perform a given task) are often limited. This is the case, for example, of battery powered nodes, which have limited energy amounts. ii) sensors may provide information that is not unique but can be generated by set of different objects which for example are capable to sense the same physical measure of the same geographical. iii) the number of nodes in the IoT is quickly overcoming the number of hosts in the 'traditional' Internet and most of these have a low reliability due mostly to the mobility and energy. This entails for a new paradigm of communication according to which objects coordinate with the other objects in groups and provide a unified service to the external world (the application that requires the service), with the intent to distribute the load of the requested services according to specific community defined rules, which could be: lifetime extension, QoS (Quality of Service) maximization, reward maximization, or others. It is evident that an appropriate coordination of objects' resources utilization would consistently improve their performance. This foreword is necessary to introduce this thesis, which is defined as follows. Task allocation in the IoT: given the IoT paradigm and the requirements of IoT applications, the nodes involved in the execution of the same application should cooperate to reach the optimal allocation of tasks among them. They should execute tasks to reach the global application target and to satisfy the relevant requirements while optimizing the network performance in terms of resources used. This issue should be continuously addressed to dynamically adapt the system to changes in terms of application requirements and network topology