Multi-Connection Scheduling with Fair Resource Management for Scalable Bluetooth Low Energy Networks

초록

Bluetooth Low Energy (BLE) is a key wireless technology for the Internet of Things (IoT), supporting low-power, multi-device communication. However, managing and scheduling multiple BLE connections under tight timing constraints remains a significant challenge. The Bluetooth specification lacks explicit strategies for efficient multi-device coordination, often leading to unfair resource allocation and connection event interruptions due to resource overlap among peripherals. To address this problem, we propose EMBLEM, a novel scheduling scheme that treats multiple connections as a unified resource list and dynamically adjusts allocations based on traffic intensity. By leveraging event preemption and connection subrating, EMBLEM enables flexible resource allocation that accommodates more peripherals while ensuring fairness. To further improve efficiency, we introduce a bitmap-based state tree for memory-efficient resource tracking, and also connection event extension to reduce resource wastage and improve throughput. We evaluate EMBLEM on a 51-node testbed against state-of-the-art scheduling schemes and popular BLE stacks. Experiment results show that EMBLEM eliminates resource overlap, improves stability, and significantly enhances system fairness, throughput, and scalability under diverse operating conditions.

키워드