AN EDGE-CLOUD TASK OFFLOADING FRAMEWORK FOR HETEROGENEOUS SMART HOME IOT NETWORKS

Abstract

Modern smart homes face critical challenges in managing heterogeneous IoT devices while meeting stringent latency requirements for safety-critical applications. This paper presents an edge-cloud IoT framework for smart home automation, addressing device heterogeneity and latency constraints through two primary contributions. First, we develop an adaptive device registration protocol handling multiple IoT communication standards (Wi-Fi, Zigbee via gateways). Second, we propose a dynamic task offloading algorithm optimising computational task distribution between edge and cloud resources based on real-time conditions. Results demonstrate a 42% latency reduction for time-critical tasks compared to cloud-only approaches, as well as 98% task completion rates under normal conditions and 92% under network congestion. The framework achieves above 90% registration success rates under severe network stress (50% packet loss) and maintains linear scalability from 10 to 50 devices with graceful latency growth (78-115ms for Wi-Fi, 95-135ms for Zigbee). The system supports scalable deployment for smart homes with up to 50 connected devices while maintaining quality-of-service guarantees for safety-critical applications.