Optimal Deployment and Orchestration of Multiple Service Function Chains in 5G Networks

This paper addresses the challenge of deploying multiple service function chains (SFCs) in 5G networks. We propose a new network model for network function virtualization (NFV) network traffic that captures the dynamically changing nature of service requests and the availability of network resources. The deployment of multiple SFCs is then formulated as a constrained optimal resource allocation problem for dynamical systems, and its approximate solution is obtained by using a constrained least-squares method. Specifically, the proposed approach treats each NFV node as a discrete-time nonlinear dynamical system and seeks the suboptimal deployment of virtual network functions (VNFs) while minimizing the overall cost function for all NFV nodes. The cost function considers both the cost for service providers and quality of service (QoS) for users. Examples are provided to illustrate the proposed new network model. Due to the analytical nature of the obtained constrained least-squares solution, we expect this algorithm to be more efficient in addressing the service requests in large-scale 5G networks.