Edge Computing Architectures for Real-Time Processing in Internet of Things Applications

Abstract

The use of edge computing architectures in real-time processing of IoT-related applications is becoming more critical especially in situations where cloud computing is limited due to high Latency, bandwidth, and scale of application. As the number of IoT devices keeps growing exponentially, cloud systems cannot effectively process the growing data volume. Edge computing can solve these problems by computing data as near as possible to the origin of the data, and enhancing real-time behavior. The paper will examine some of the edge computing architectures that are customized to IoT use in smart cities, industrial automation, and healthcare monitoring. It contains a comparative study of edge computing models regarding Latency, throughput, energy consumption, and reliability. The statistical results indicate that edge computing cuts down the Latency by 5-6 times (50 ms in edge computing and 250ms in cloud-based systems), raises throughput by 2.5-3 times (2000ops/sec in edge computing and 800ops/sec in cloud-based systems) and a 25-35 % cut in energy consumption (1.2J in edge computing and 3.5J in cloud-based systems). Besides, the edge computing model is more accurate with an accuracy of 96.3 and an F1-score of 94.5 when compared to the cloud systems. These findings indicate the great performance enhancement of edge computing in the IoT processes, which guarantees rapid decision-making, energy-saving, and scalability. The paper concludes that edge computing has a potential solution to the IoT systems that need real-time processing, and some more research is required before the hybrid edge-cloud can be optimized to achieve even higher flexibility and performance.