Network function virtualization (NFV) is designed to reduce high cost of the hardware deployment and maintenance, which will be of great importance in the future resource management. NFV makes it possible for general servers to realize fast deployment of network functions, namely virtual network function (VNF), to achieve on-demand requests. The service request is generally composed of a sequence of VNFs, which is defined as a service chain. Although one node (server) only run one VNF at a time, the VNF could be shared with several VNF chains. How to deploy VNF and share VNF to achieve high usage of both computing and bandwidth resource is a challenge. In this paper, we present the VNF request model and the formal definition of the VNF deployment problem. Then, we investigate the VNF reuse and deployment problem, and propose a greedy strategy to deploy VNFs, so that the system could effectively offer a reliable VNF service. Simulations show that this VNF deployment schema could effectively improve the resource usage compared with other classic algorithms.
CITATION STYLE
Xu, D., Li, Y., Yin, M., Li, X., Li, H., & Qian, Z. (2017). A reliable resource scheduling for network function virtualization. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10658 LNCS, pp. 251–260). Springer Verlag. https://doi.org/10.1007/978-3-319-72395-2_24
Mendeley helps you to discover research relevant for your work.