A Mini Review on Edge Computing Devices for IoT Systems

Abstract

Edge computing is a transformational distributed computing paradigm that brings computers and data storage closer to IoT data sources, overcoming critical limitations of cloud-centric approaches such as excessive latency, bandwidth constraints, privacy concerns, and energy inefficiency. This review examines the fundamental requirements of edge computing devices for IoT, exploring essential characteristics including processing capabilities (through microcontrollers, microprocessors, and AI accelerators like GPUs, TPUs, and VPUs), energy efficiency mechanisms, ultra-low latency solutions, connectivity options (LoRa, Wi-Fi, BLE, 5G), and robust security frameworks. Current challenges in designing edge computing for IoT applications include benchmarking difficulties due to a lack of standardized metrics, severe resource constraints, security vulnerabilities in distributed networks, scalability issues in dynamic environments, and limited real-time adaptability with pre-trained models. Emerging technologies such as federated learning, 6G connectivity, TinyML innovations, and sustainable design approaches show significant promise for improving edge computing efficiency. The findings reveal that edge computing has profoundly impacted IoT development by enabling real-time processing and decision-making across industries, enhancing data privacy, improving energy efficiency, and facilitating AI integration into resource-constrained environments, while continuing research focuses on addressing persistent challenges in scalability, security, standardization, and sustainable deployment.