Latency and Cost-Aware Deployment of Dynamic Service Function Chains in 5G Networks

Efficient deployment of a service function chain (SFC) on virtual network functions (VNFs) and their mapping to virtual machines (VMs) in 5G networks is highly important to reduce user application service latency as well as to minimize the cost of VM migration. Existing works in the literature have either considered service latency or migration cost while deploying an SFC on VMs. In this paper, the problem of dynamically mapping VNFs running a SFC to different virtual machines in a 5G network has been formulated as a multi-objective linear Programming (MOLP) problem. The developed optimization model, namely Trade-Lcm, minimizes user application service latency while reducing the cost of migrating virtual network functions associated with an SFC. The numerical performance analysis results demonstrate a significant improvement in minimizing service latency of user applications and cost of VM migration as high as 30% and 10%, respectively, compared to the state-of-the-art work.