Optimization of Flow Allocation in Asynchronous Deterministic 5G Transport Networks by Leveraging Data Analytics

Abstract

Time-Sensitive Networking (TSN) and Deterministic Networking (DetNet) technologies are increasingly recognized as key levers of the future 5G transport networks (TNs) due to their capabilities for providing deterministic Quality-of-Service and enabling the coexistence of critical and best-effort services. Addi- tionally, they rely on programmable and cost-effective Ethernet- based forwarding planes. This article addresses the flow alloca- tion problem in 5G backhaul networks realized as asynchronous TSN networks, whose building block is the Asynchronous Traffic Shaper. We propose an offline solution, dubbed “Next Generation Transport Network Optimizer” (NEPTUNO), that combines ex- act optimization methods and heuristic techniques and leverages data analytics to solve the flow allocation problem. NEPTUNO aims to maximize the flow acceptance ratio while guaranteeing the deterministic Quality-of-Service requirements of the critical flows. We carried out a performance evaluation of NEPTUNO regarding the degree of optimality, execution time, and flow rejection ratio. Furthermore, we compare NEPTUNO with a novel online baseline solution for two different optimization goals. Online methods compute the flow’s allocation configuration right after the flow arrives at the network, whereas offline solutions like NEPTUNO compute a long-term configuration allocation for the whole network. Our results highlight the potential of data analytics for the self-optimization of the future 5G TNs.