• Internet of Things
• QoS for IoT
• CoAP
• embedded
• thing as a service

The Internet of Things (IoT) is rapidly becoming reality, the cut off prices as well as the advancement in the consumer electronic field are the two main training factor. For this reason, new application scenarios are designed every days and then new challenges that must be addressed. In the future we will be surrounded by many smart devices, which will sense and act on the physical environment. Such number of smart devices will be the building block for a plethora of new smart applications which will provide to end user new enhanced service. In this context, the Quality of Service (QoS) has been recognized as a non functional key requirement for the success of the IoT. In fact, in the future IoT, we will have different applications each one with different QoS requirements, which will need to interact with a finite set of smart device each one with its QoS capabilities. Such mapping between requested and offered QoS must be managed in order to satisfy the end users.

The work of this thesis focus on how to provide QoS for IoT in a cross-layer manner. In other words, our main goal is to provide QoS support that, on one hand, helps the back-end architecture to manage a wide set of IoT applications, each one with its QoS requirements, while, on the other hand, enhances the access network by adding QoS capabilities on top of smart devices.
We analyzed existing QoS framework and, based on the status of the art, we derive a novel model specifically tailored for IoT systems. Then we define the procedures needed to negotiate the desired QoS level and to enforce the negotiated QoS. In particular we take care of the Thing selection problem which is raised whenever more than one thing can be exploited to obtain a certain service.

Finally we considered the access network by providing different solutions to handle QoS with different grain scale. We proposed a totally transparent solution which exploits virtualization and proxying techniques to differentiate between different class of client and provide a class based prioritization schema. Then we went further by designing a QoS framework directly on top of a standard IoT protocol called Constrained Application Protocol (CoAP). We designed the QoS support to enhance the Observing paradigm
which is of paramount importance especially if we consider industrial applications which might benefit from a certain level of QoS assurances.