Performance and trust driven scheduling in heterogeneous fog networks

Abstract

Fog computing delivers cloud-like facilities (using cloudlets or fog devices) at the network edge. It provides computing services with ultra-low-latency access, yielding highly responsive computing services to user requests. This thesis proposes several scheduling algorithms for heterogeneous fog networks. The first work proposes scheduling algorithms for hierarchical fog-cloud architectures with multiple layers of fog devices between the users and the cloud data center. We formulate our research problem by jointly minimizing the completion time and energy consumption on an n-tiered fog-cloud architecture. To reduce the completion times of the jobs, we propose two hierarchical scheduling algorithms: F iF SA (Hierarchical First Fog Scheduling Algorithm) and EF SA (Hierarchical Elected Fog Scheduling Algorithm). F iF SA and EF SA offer an improvement of 19% to 70% in completion times and 42% to 72% in system cost over comparable algorithms, respectively. In the second work, we propose a real-time heterogeneous hierarchical scheduling algorithm called RT H2S. We consider a hierarchical model for fog nodes, with nodes at higher tiers having greater computational capacity than nodes at lower tiers, though with greater latency from data generation sources. Typically, in terms of computation needs, we consider three kinds of jobs: small (S), medium (M), and large (L). Further, we consider three types of job deadlines: tight (T), loose (L), and no deadline (N). We also consider Microsoft Azure-based costs for the proposed algorithm. A real-life workload demonstrates the algorithm’s performance using a simulation and prototype testbed. We observe that RT H2S offers better real-time results in higher Success Ratios and reduced Monetary Costs. The third work presents a technique to schedule real-time jobs on trustworthy fog devices. Fog devices are generally susceptible to privacy, security, and trust issues. We propose RT − T ADS (Real Time-Trust Aware Dynamic Scheduling), a trust-based scheduling model that considers the tasks’ deadlines and privacy requirements. We offer a trust computation model to compute the trustworthiness of fog devices. Tasks submitted by users are tagged as private, semi-private, and public. Likewise, fog devices are classified into extremely high, high, normal, low, and untrusted. Results indicate that the proposed trust-based scheduling algorithm can schedule tasks with stricter latency requirements on the appropriate fog devices, while maintaining a higher task success ratio. In the fourth work, a case study of a fog-based fire evacuation framework is presented. We propose F AF CA (Fog Assisted Fire Control Algorithm) that guides the best path for the evacuees present in the building after considering various parameters, such as exit capacity, distance to exits, and distribution of evacuees. The intuition is that due to a lower communication latency between the user & the fog nodes, the processing is done quickly, resulting in rapid evacuation decisions. Simulation results reveal that the F AF CA performs better compared to baseline and cdc-only schemes. This thesis also provides a detailed description of the fog computing paradigm. It summarises the state-of-the-art work with respect to the above-discussed four problems. Finally, this thesis analyzes future research directions in this field.