Sensing and Actuation as a Service, a device-centric paradigm for the IoT: analysis, design and case studies

Abstract

The huge and steady growth in the number of distributed devices connected to the global network as a so-called Internet of Things (IoT) calls for infrastructure management techniques able to deal with this overwhelming complexity, especially in light of the growing impact of the sharing economy and the role played by the so-called "long tail". In this context, the as-a-Service approach provides well investigated mechanisms for infrastructure and service provisioning, and an interesting challenge lies in evaluating its application to the instantiation and lifecycle management of a dynamic, possibly virtualised, infrastructure of sensing (and actuation) resources. Apart from the flexibility of renting this kind of resources according to the utility model, by providing access at the lowest level where possible, the research path undertaken in this dissertation is meant to provide even bigger scope to the scenarios under consideration, from a platform for opportunistic and cooperative mobile crowd-sensing to a model for Software-Defined Smart Cities, where dynamic reconfiguration of the wiring among Things ultimately enables wide-area feedback control loops on demand. Yet none of these outcomes, and in particular shaping the surrounding environment, may be achievable without means to interact bidirectionally, and with the greatest control that may be exerted, with physical systems through remote sensors and actuators at one's own fingertips. The premise then lies in engaging the research from a "device-centric" perspective. The infrastructure and platform takes on IoT, the Stack4Things framework as well as other results are then outcomes of this approach.

Analytical descriptions of the scenarios, details of the architectures and investigated case studies therefore are here provided, while also reporting and highlighting design choices.