Scheduling bag-of-task jobs with security requirements and dual-criteria partial computations in a fog – cloud system

Abstract

Internet of Things (IoT) applications have become integral to daily life, but they often impose stringent security and timing constraints. Fog computing, a modern computing paradigm, extends the capabilities of cloud resources to the edge of the network, offering potential solutions to these challenges. In these collaborative computing environments, workloads are composed of tasks with varying security requirements. Consequently, implementing security-aware scheduling schemes is crucial to ensure the proper execution of applications and to achieve the desired Quality of Service (QoS). This study explores security-conscious scheduling methodologies specifically designed for Bag-of-Tasks (BoT) applications within a fog-cloud computing framework. Fog resources are suitable for processing BoT tasks with specific security requirements, referred to as “fog jobs”. Conversely, BoT applications without specific security requirements, referred to as “cloud jobs”, can be processed on either cloud or fog resources. To expedite the execution of fog jobs, a method that utilizes partial computations for cloud job tasks is employed. This study provides valuable insights into the effects of approximations on system performance through simulation-based evaluations. It considers partial computations based on delay levels in relation to service demands and the size of cloud jobs. This approach enhances system performance with only a minor trade-off in results accuracy.