Latency-Aware and QoS-Optimized Service Function Chaining in Multi-Tier SDN-Enabled 5G Network Architectures

Abstract

The Service Function Chaining (SFC) is essential to service delivery of end-to-end Quality of Service (QoS) within the 5G network by dynamically orchestrating a group of Virtual Network Function (VNF). Nevertheless, the newfound 5G applications latency sensitivity in conjunction with the dynamicism of traffic loads presents very serious problems to effective service provisioning. The current paper describes a latency-aware and Quality of Service (QoS)-optimized SFC architecture that can be used in multi-tier Software-Defined Networking (SDN)-enabled 5G systems. The proposed architecture will rely on a hierarchical SDN control plane, which includes the edge, fog and core controllers, to facilitate a distributed management of network resources. In order to efficiently perform VNF placement and service path selection on-demand, we include a Deep Reinforcement Learning (DRL) agent which is trained with information about the state of the network in terms of the latencies, bandwidth and QoS attributes. The set of experiments on a multi-tier SDN setting proves that our mechanism has the ability to considerably lessen the end-to-end latency and QoS contravention and optimize the overall use of network resources better. The proposed model contains up to 43 percent reduction in latency and more than 97 percent QoS compliance in various traffic conditions when compared to conventional centralized and greedy SFC algorithms. These outcomes validate the viability and feasibility of the suggested solution and it is a very possible course of action to implement the given solution into the 5G and beyond-5G network infrastructures of the future.