A High-Throughput Scheduler based on Multipath-State Machine in Wireless Networks

Multipath transmission is an important solution in 6G wireless networks to improve communication resource convergence among heterogeneous networks. The appearance of out-of-order (OFO) packets is a problem in multipath transmission, especially in heterogeneous wireless networks, where network states fluctuate frequently. Existing schedulers attempt to solve the OFO problem by designing an "in-order" strategy. However, these "in-order" strategies lose their roles in jittered networks, resulting in poor throughput performance. In this study, we propose a novel packet scheduling strategy, called a multipath-state machine (MSM) scheduler. MSM prioritizes improving throughput utilization over ensuring packets arrive in order. Based on whether links are fully utilized, MSM divides link utilization states into 5 phases, including establishment, idle, busy, congestion, and risk. MSM prefers underutilized links. Furthermore, MSM chooses the one with the shortest transmission delay when multiple paths are in the same utilization state. Experiments show that MSM provides a more stable and higher bandwidth utilization than typical algorithms in various networks with heterogeneous characteristics such as delay, bandwidth, and jitter.