Service chain performance optimization based on middlebox deployment

Abstract

The Service Function Chain (SFC) is a mechanism which transmit packets through different services in order. With Software Defined Networking (SDN) and Network Function Virtualization (NFV), network services can be flexible deployed on the middleboxes. However, end-to-end transmission delay of the SFC will influence the time delay in a network, which is a key indicator to the performance of a network. One way to decrease transmission delay is to optimize the deployed sequence of service middleboxes. In this paper, we put forward a Middlebox Placement Optimization (MPO) algorithm. Different from the previous works, the MPO algorithm fully utilizes the flexibility and dynamicity provided by SDN and NFV, and can dynamically deploy the sequence of services in the SFC to adapt real-time changing service characters. Simulation results show that MPO can reduce the end-to-end delay by an average of 30 percent comparing with the traditional First Construction First Placement (FCFP) algorithm.