Service function chaining to support ultra-low latency communication in NFV

Abstract

By exploiting the benefits of virtualization and cloud computing technologies, Network Function Virtualization enables networks to be more flexible, manageable, and scalable. Ultra-low latency applications in 5G, 6G and IoT demand very low latency and assured QoS. With limited network resources, network providers must develop an effective strategy to support ultra-low latency applications. We propose a novel Service Function Chaining algorithm in this paper with the goal of minimizing latency and optimizing physical resource allocation for ultra-low latency applications while having the minimum possible negative effects on other applications. We prioritize ultra-low latency traffic flows and enable them to optimize their provisioning paths by using reserved physical resources (bandwidth, CPU, and memory). We provide a mathematical model for the SFC embedding problem in the form of an Integer Linear Programming optimization model that takes QoS constraints into account (related to latency and consumption of links and servers). We present a heuristic algorithm for obtaining near-optimal solutions with the smallest possible optimality gap and execution time, allowing it to be applied to real-world network topologies. The performance evaluations show that our proposed algorithms effectively provide better results for ultra-low latency applications in terms of end-to-end delay (up to 20 percent), bandwidth utilization (up to 27 percent) and SFC acceptance rate (up to 10 percent) compared to the existing algorithms.