Preemption-Aware Rank Offloading Scheduling for Latency Critical Communications in 5G Networks

Abstract

This paper introduces a preemptive rank offloading scheduling framework for joint ultra-reliable low-latency communications (URLLC) and enhanced mobile broadband (eMBB) traffic in 5G new radio (NR). Proposed scheduler dynamically adapts the overall system optimization among the network-centric ergodic capacity and the user-centric URLLC one-way latency, based on the instantaneous traffic and radio resources availability. The spatial degrees of freedom, offered by the transmit antenna array, are fully exploited to maximize the overall spectral efficiency. However, when URLLC traffic buffering is foreseen, proposed scheduler immediately enforces scheduling pending URLLC payloads through preemption-aware subspace projection. Compared to the state-of-the-art schedulers from industry and academia, proposed scheduler framework shows significant scheduling flexibility in terms of the overall ergodic capacity and URLLC latency performance. The presented results therefore offer valuable insights of how to most efficiently multiplex joint URLLC-eMBB traffic over the 5G NR spectrum.