A Simulation-based Performance Evaluation of Heuristics for Dew Computing

Abstract

The evolution of smartphones allows the continuous exploitation of computing resources. This increasingly applies also to distributed environments as exemplified through utilization of network router loads in edge computing and fog computing. Combining cloud computing and mobile smart devices in dew computing contexts enables new techniques for resource utilization, data collection and data processing. However, new challenges regarding job scheduling arise. Smartphones may be used in ad-hoc networks in this context, but their heterogeneity and energy usage must be considered. We propose novel heuristics for performance measuring of distributed computing systems integrated with mobile devices and compare them with previous heuristics in a simulation environment. Our results show an overall improvement in job completion and load balancing metrics compared to previous approaches. They highlight the usefulness of pursuing this research stream for aiming at industrial implementation and evaluation.