A Novel NFV Schedule Optimization Approach with Sensitivity to Packets Dropping Positions

Abstract

Network Function Virtualization (NFV) is a promising approach for providing efficient and flexible network functions deployment, which also reduces the cost in the network operations compared to traditional method using dedicated hardware. Multiple virtual network functions can be chained together working as service chain to provide a specific service. When multiple network functions share computing resources, how to efficiently schedule them dynamically is a great challenge. In this paper, we first investigate the issue that a single packet drop occurs at different positions of a service chain may have different impact on the performance. In light of this, a packets dropping cost model is carefully defined based on the NF positions in a service chain. Then, according to linear programming, a Minimizing Dropping Cost (MDC) approach is proposed to optimally schedule vNFs. Simulation results show that the proposed algorithm can reduce the total incurred cost by up to 60% compared to the existing common schedule approach.