RETALIN: A Queue Aware Uplink Scheduling Scheme for Reducing Scheduling Signaling Overhead in 5G NR

Abstract

Flexible numerology in 5G New Radio (NR) helps to reduce End-to-End (E2E) delay by supporting different slot duration options. But, higher numerologies can increase signaling overhead (in terms of Scheduling Requests (SRs) in uplink (UL)) for applications which are UL heavy in nature; as a consequence, the E2E delay of such applications could be impacted. This sudden increase in SRs in UL happens when the timing of the Constant-Bit-Rate (CBR) UL transmission is not properly aligned with the slot duration of the numerology employed and aggressive emptying of the buffered data at the User Equipments (UEs) by the UL scheduler. In this work, to minimize SRs while meeting the latency requirements of the UL heavy applications like vehicular applications, we propose a novel UL MAC scheduling algorithm named RETALIN. The proposed RETALIN is a queue-aware radio resource scheduler that estimates the probability of SR with respect to each UE (vehicle) and anatomizes the transient queue behavior by controlling backlogs of the UEs in order to reduce SRs and thereby mitigating an adverse impact of numerology in the UL traffic while keeping a bound on E2E delay of the vehicular applications. Extensive NS-3 simulations are performed to evaluate the performance of the proposed RETALIN scheme with mobility traces taken from Simulation of Urban Mobility (SUMO) using OpenStreetMap. Simulation results show that the proposed RETALIN scheduling scheme significantly reduces link delay for different numerologies when compared with a state-of-art QoS scheduler. Further, RETALIN increases Packet Delivery Ratio (PDR) while reducing signaling overhead due to SRs for UL scheduling. In the case of a High Definition Map (HD Map) vehicular application, RETALIN assists in increasing the Offloading Success Rate (OSR) over the QoS scheduler.