Dynamic Orchestration of Edge-Cloud Workloads for Real-Time Industrial IoT Applications

Abstract

The paper focuses on addressing the critical challenges involved in managing computational workloads within Industrial Internet of Things (IIoT) environments. IIoT applications often generate vast amounts of data requiring timely processing to meet stringent real-time performance requirements. Traditional centralized cloud solutions struggle to satisfy these low-latency demands due to inherent network delays and bandwidth limitations. Conversely, edge computing nodes offer reduced latency by processing data closer to its source but are constrained by limited computational resources. To overcome these limitations, the paper proposes a dynamic orchestration framework designed to intelligently and efficiently distribute computational tasks between edge devices and cloud servers. This framework continuously monitors the state of resources, network conditions, and task requirements, allowing it to make informed, adaptive decisions regarding where each task should be executed. By dynamically balancing the workload across the edge-cloud continuum, the framework aims to optimize resource utilization, minimize processing delays, and ensure that real-time constraints are consistently met. Moreover, it is designed with scalability in mind, enabling the system to gracefully handle increasing volumes of data and a growing number of devices without sacrificing performance or reliability.