5G RAN Slicing with Load Balanced Handovers

Abstract

With increasing density of small cells in modern multi-cell deployments, a given user can have multiple options for its serving cell. The serving cell for each user must be carefully chosen such that the user achieves reasonably high channel quality from it, and the load on each cell is well balanced. It is relatively straightforward to reason about this without slicing, where all users can share a global load balancing criteria set by the network operator. In this paper, we identify the unique challenges that arise when balancing load in a multi-cell setting with 5G slicing, where users are grouped into slices, and each slice has its own optimization criteria, resource quota, and demand distributions, making it hard to even define which cells are overloaded vs underloaded. We address these challenges through our system, RadioWeaver, that co-designs load balancing with dynamic quota allocation for each slice and each cell. RadioWeaver defines a novel global load balancing criteria across slices, that allows it to easily determine which cells are overloaded despite the fact that different slices optimize for different criteria. Our evaluation, using large-scale trace-driven simulations and a small-scale OpenRAN testbed, show how RadioWeaver achieves 16-365% better performance when compared to several baselines.