Optimal virtual PON slicing to support ultra-low latency mesh traffic pattern in MEC-based Cloud-RAN

TL;DR

This paper presents a mixed analytical-iterative model for optimal virtual PON slice allocation to support ultra-low latency mesh traffic in MEC-based Cloud-RAN, enabling real-time network slicing for 5G and beyond.

Contribution

It introduces a novel mixed model for dynamic virtual PON slicing that ensures ultra-low latency mesh connectivity in MEC-based Cloud-RAN networks.

Findings

Provides a real-time compatible algorithm for virtual PON slice allocation.

Achieves ultra-low end-to-end latency for mesh traffic.

Supports high reliability in 5G and beyond networks.

Abstract

As progressive densification of cells, deployment of Cloud-RAN and \ac{MEC} are coming into reality to support the ultra-low latency with high reliability in 5G and beyond, it generates mesh traffic pattern across fronthaul network. This led to evolution of PON architectural enhancements with virtualization in order to support such mesh traffic pattern. However, allocation of virtual PON slices dynamically over such mesh-PON based fronthaul transport is becoming a research challenge. In this paper, we provide a mixed analytical-iterative model to compute optimal virtual PON slice allocation, providing mesh access connectivity with ultra-low end-to-end latency in next-generation MEC-based Cloud-RAN. Our proposed method can compute optimal virtual PON slice allocation in timescales compatible with real-time or near real-time operations.