Novel Micro-Burst Detection Method in Access/Edge Networks Using Periodic IoT Messages

Abstract

Real-time access edge computing demands access networks to connect a lot of IoT devices to edge servers with ultra-low latency. A lot of connected devices will cause the data bursts and sudden increase of delay (i.e. micro-burst) in the upstream queues in gateways. This study proposes a novel method for detecting such bursts to ensure low-latency communication. The proposed method leverages the periodic messaging behavior of certain IoT devices. It detects bursts based on the variation in the arrival times of messages transmitted from already deployed multiple IoT devices. This approach enables monitoring of delays and detecting bursts without the need for additional probes or modifications to gateways. The study validates the detection accuracy and feasibility through numerical analysis and simulations.