Importance Measures for NFV Data Center: An Availability Evaluation

Abstract

Data center infrastructures need to provide high availability of their services. Unexpected spikes of downtime in data centers lead to financial losses. Besides, there are intangible costs such as damaged reputation, low employee satisfaction, and reduced customer retention. In this context, Network Function Virtualization (NFV) emerged as a paradigm that assists data centers in becoming more dynamic and flexible. This paper presents an availability evaluation and importance analysis under the redundancy of NFV in data centers. The NFV data center component importance is represented by using Reliability Block Diagrams (RBD). The proposed models have identified the availability importance and critical devices in an NFV data center. This research also suggests alternatives for device redundancy to reach higher availability and proposes a new importance measure for analyzing the impact of NFV factors on data center availability. The measure can evaluate the degree of the impact of a failure on the data center, therefore, help to identify the factors with a substantial impact on data center availability.