Integrating edge computing with advanced cloud computing: A paradigm shift for IoT applications

In this paper we explore the coupling of edge computing with the superset advanced cloud computing and its applicability to the area of the Internet of Things (IoT), and we explore a paradigm shift in how data is processed and managed by connected devices. We prove experimentally that we achieve significant reductions in latency (86% on average), bandwidth optimization (58% data reduction on average), and energy efficiency (67% decrease in transmission related power consumption) through extensive experimentation and analysis. With the proposed integrated architecture, we show that we can scale up with minimal cost: we can accommodate a 500% increase in the number of connected devices while only incurring a 35% increase in cloud infrastructure. There are challenges in security, standardization and ultra-scale deployment. In this thesis we show that edge-cloud integration provides substantially higher benefits over conventional cloud-only and cloud-edge integration methods for most of the IoT scenarios investigated in this thesis, but adds new complexities that remain unexplored and need further research and development. These results in contribute to the developing framework of the IoT infrastructures, informing researchers and practitioners designing large scale, distributed systems, often with centralized control.