On Analyzing the Trade-Off Between Over-Commitment Ratio and Quality of Service in NFV Datacenter
Network Function Virtualization (NFV) is one of the key technologies in 5G. It inherits virtualization technology in cloud computing and promises to bring many benefits for industry such as saving energy and reducing capital expenditure. Resource over-commitment is a significant technique of virtualization to fully utilize the resources of a cloud datacenter. Deploying VNFs with difference over-commitment ratios in an NFV datacenter leads to different results of the number of servers used and the order of VNFs placed in each physical server. This also affects Quality of Service (QoS) and energy consumption in the NFV datacenter. However, to the best of our knowledge, there has been no study to find exactly how much resource over-commitment is sufficient to meet the QoS while reducing the energy used in an NFV datacenter. In this paper, we analyze and evaluate the effect of CPU over-commitment ratio in VNF placement problem while considering the QoS and energy efficiency for NFV datacenters. After exhausting simulations, we have found out a proper value for CPU over-commitment ratio for NFV datacenters. By employing that ratio, an NFV datacenter could reduce up to 16.8% total power consumption compared with the others not using the over-provisioning technique.
KeywordsVNF placement Over-commitment ratio Saving energy OpenStack Network function virtualization
This research was supported in part by Korean government, under by AI Graduate School Support Program (No. 2019-0-00421) supervised by the Ministry of Science and ICT (MSIT) and ICT Consilience Creative program (IITP-2019-2015-0-00742) supervised by the Institute of Information & Communications Technology Planning & Evaluation (IITP), respectively.
- 2.ETSI white paper, Network Function Virtualization: Architectural Framework (2013). http://www.etsi.org/deliver/etsi_gs/nfv/001_099/002/01.01.01_60/gs_nfv002v010101p.pdf
- 3.Beloglazov, A., Buyya, R.: OpenStack neat: a framework for dynamic consolidation of virtual machines in OpenStack clouds - a blueprint. Cloud Computing and Distributed Systems (CLOUDS) Laboratory (2012)Google Scholar
- 4.Al-Shabibi, A: CORD: Central Office Re-architected as a Datacenter, A Whitepaper by ON.LAB, AT&T, ONOS and PMC, OpenStack Summit (2015)Google Scholar
- 5.Davis, D.M.: Demystifying CPU Ready (% RDY) as a Performance Metric, Dell white paper (2012)Google Scholar
- 10.Rani, A., Peddoju, S.K.: A workload-aware VM placement algorithm for performance improvement and energy efficiency in OpenStack Cloud, In: ICCCA (2017)Google Scholar
- 11.Panigrahy, R., Talwar, K., Uyeda, L., Wieder, U.: Heuristics for vector bin packing (2011). http://research.microsoft.com
- 13.ur Rahman, H., Wang, G., Chen, J., Jiang, H.: Performance evaluation of hypervisors and the effect of virtual CPU on performance. In: IEEE SmartWorld (2018)Google Scholar
- 14.Pham, C., Tran, N.H., Ren, S., Saad, W., Hong, C.S.: Traffic-aware and energy-efficient vNF placement for service chaining: joint sampling and matching approach. IEEE Trans. Serv. Comput. (2018) Google Scholar
- 15.Zhang, X., Wu, C., Li, Z., Lau, F.C.: Proactive VNF provisioning with multi-timescale cloud resources: fusing online learning and online optimization. In: IEEE INFOCOM (2017)Google Scholar
- 16.Marotta, A., Kassler, A.: A power efficient and robust virtual network functions placement problem. In: 28th International Teletraffic Congress (2016)Google Scholar
- 18.Ranjana, R., Radha, S., Raja. J.: Performance study of resource aware energy efficient VM placement algorithm. In: IEEE WiSPNET (2016)Google Scholar
- 19.ENERGY STAR Power and Performance Data Sheet. http://www.dell.com/downloads/global/products/pedge/en/Dell-PowerEdge-R620-750W-E5-2620-40-Family-Data-Sheet.pdf