Multi-Cell, Multi-Channel Scheduling with Probabilistic Per-Packet Real-Time Guarantee

For mission-critical sensing and control applications such as those to be enabled by 5G Ultra-Reliable, Low-Latency Communications (URLLC), it is critical to ensure the communication quality of individual packets. Prior studies have considered Probabilistic Per-packet Real-time Communications (PPRC) guarantees for single-cell, single-channel networks with implicit deadline constraints, but they have not considered real-world complexities such as inter-cell interference and multiple communication channels. Towards ensuring PPRC in multi-cell, multi-channel wireless networks, we propose a real-time scheduling algorithm based on \emph{local-deadline-partition (LDP)}. The LDP algorithm is suitable for distributed implementation, and it ensures probabilistic per-packet real-time guarantee for multi-cell, multi-channel networks with general deadline constraints. We also address the associated challenge of the schedulability test of PPRC traffic. In particular, we propose the concept of \emph{feasible set} and identify a closed-form sufficient condition for the schedulability of PPRC traffic. We propose a distributed algorithm for the schedulability test, and the algorithm includes a procedure for finding the minimum sum work density of feasible sets which is of interest by itself. We also identify a necessary condition for the schedulability of PPRC traffic, and use numerical studies to understand a lower bound on the approximation ratio of the LDP algorithm. We experimentally study the properties of the LDP algorithm and observe that the PPRC traffic supportable by the LDP algorithm is significantly higher than that of a state-of-the-art algorithm.