Towards delay-optimized and resource-efficient network function dynamic deployment for VNF service chaining

By decoupling virtualized network functions from the dedicated network equipment on which they run, Network Function Virtualization (NFV) has brought a flexible and economical way to support the complex communication demands of different applications. Virtualized Network Functions (VNFs) can be dispatched and deployed as instances of plain software on or near the communication paths of applications to establish Service Function Chains (SFCs), so as to provide special packet processing operations beyond simple packet forwarding. However, it is still a great challenge to dynamically place appropriate network functions at suitable locations so as to improve the efficiency of establishing SFCs and optimize network resource utilization. In this paper, the mechanism of dynamically deploying customized network functions via NFV is proposed. By predicting the future popularities of applications to switches, it adaptively places most of the appropriate network functions in corresponding forwarding equipment before they are massively requested. The serious latency and extra resource consumption caused by real-timely dispatching and deploying most of the requested network functions will be avoided. Then, the approach of Ant Colony Optimization (ACO) inspired multi-switch cooperative network function providing is devised. By cooperating multiple forwarding equipment on the packet transmission path, it makes full use of the already placed network functions to support packet processing operations in time with the cost and delay factors jointly considered. Simulation results show that the proposed mechanism has significant improvements in time overhead and resource utilization compared with the current state of the art. Specifically, our mechanism is capable of improving the service delay, the function utilization ratio, and the SFC adjustment efficiency by about 14%, 10% and 12% respectively, compared with corresponding work.