Lightweight Virtualization in Edge Computing for Internet of Things

Lightweight virtualization technologies have revolutionized the world of software development by introducing flexibility and innovation to the hardware and software virtualization domain. In particular, container-based virtualization has spread widely across the industry positively impacting the development of new-generation applications. Although the advantages introduced by these emerging software solutions have been widely acknowledged in cloud computing environments, more recent enhancements have also led to an extensive use of such technologies in additional contexts. Internet of Things (IoT) and edge computing have started to benefit from the new virtualization paradigms. However, not all of the valuable potentialities introduced by such technologies have been widely explored. The main goal of this dissertation is to fill the gap in this respect and to show how container technologies can be easily turned into a powerful tool to be used in a range of pervasive and ubiquitous environments. Specifically, we focus on the possibility of devising an effective use of lightweight virtualization on devices characterized by limited computational resources – such as Single-Board Computers (SBCs) – in order to allow an effective redesign of key entities for specific IoT and edge computing deployments. The contributions of this dissertation can be divided into four categories. First, we describe how edge computing can empower the IoT by enabling more flexible and versatile network architectures, allowing the delivery of ever more performing services. Second, in view of the further challenges introduced by IoT edge computing scenarios, we envisage the use of lightweight virtualization to overcome such challenges, either in full or in part. Next, we look at four distinct use cases and propose to take advantage of certain characteristics of container-based virtualization for overcoming some of the limitations which are typical of these four representative IoT domains. Finally, based on the research investigations performed, we outline to what extent container-based virtualization can empower IoT and edge computing by highlighting its features and its performance weaknesses. In the four analyzed IoT and edge computing domains we characterize and validate the use of container-based virtualization through the implementation of distinct proofs of concept. Furthermore, we perform extensive empirical analyzes in order to further assess the feasibility in using this emerging technology as a key enabler of forthcoming IoT and edge computing scenarios. Such methodology proved successful in obtaining several insights, and in identifying open issues and future research challenges in the domains investigated.