Real-Time Task Offloading with Replication Under Node Churn in Volunteer Edge Computing

Abstract

Energy efficiency is a primary design objective for battery-powered IoT devices. While offloading computation-intensive tasks to edge servers has been extensively studied to mitigate power drains, relatively little attention has been paid to the long-term financial cost of commercial edge services. This article proposes Volunteer Edge, a cost-effective real-time task offloading framework that exploits underutilized computing resources of privately managed nodes to execute offloaded workloads. Unlike conventional public edge servers, volunteer edge nodes provide inexpensive computing resources but are subject to unpredictable node churn. To address this challenge, we present a dual-class task model that partitions workloads into critical and normal tasks, and selectively applies task replication to volunteer edge nodes. The framework jointly optimizes task placement, processor frequency scaling, and replication decisions using a steady-state genetic algorithm to minimize task execution cost and IoT-device energy consumption while satisfying schedulability and reliability constraints. Extensive simulations demonstrate that Volunteer Edge significantly reduces offloading cost while maintaining IoT-device energy efficiency and protecting critical tasks against volunteer node failures. Specifically, the proposed framework reduces edge rental costs by 54.0% on average compared with conventional public-edge-based offloading while maintaining reliable execution of critical real-time tasks.