Year: 2026
Type: article
Abstract: The recent explosion of event-driven applications in smart cities, industrial Internet of Things (IIoT) and healthcare monitoring systems has necessitated a desperate need of low-latency and efficient data processing architectures. Traditional cloud-focused solutions are characterized by excessive latency of communication, whereas entirely edge-based solutions are limited by limited computing and storage facilities. In order to overcome these issues, this paper proposes a hybrid edge-cloud data processing architecture of latency-sensitive event-driven applications. The proposed framework also includes a dynamic decision engine that optimally decides where to place the tasks between the edge and cloud layers based on the real-time estimation of the latency, workload characteristics, and network conditions. A mathematical latency model is developed to measure transmission, queueing and processing delays so that informed offloading decisions can be made. Moreover, an event classification system gives priority to the latency-aware tasks to be executed on the edges and assigns the workloads that are computationally intensive to the cloud. The structure is tested in a simulation environment, where there are dissimilar workloads, and different network environments. Experimental findings show that the given approach leads to massive improvements in average latency and task response time, whereas the resource utilization is also better than in traditional edge-only and cloud-only models. The suggested system offers a scalable and effective system to meet event-driven applications in real-time with strict latency requirements.
Author: Metahun Lemeon, Jinfe Regash
Institution: Addis Ababa University
Language: English
Topic: IoT and Edge/Fog Computing
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7219985256
