Analysis of Availability Model Based on Software Aging in SDN Controllers with Rejuvenation

Abstract

Deficiency of flexibility and programmability of legacy network architecture has been the concern of many networking admirers for some years. Software defined networking (SDN) is a new emerging concept to logically centralize the network control plane in order to automate the configuration of individual network elements. However, the failures of SDN controller are every large impact on the network performance and availability. There are different failure modes in SDN controller outages such as hardware and software. Unplanned downtime is mostly caused by software failure due to software aging rather than hardware failure. The aging related faults have a huge effect on the availability for software components, SDN controllers. For that reason, the work presented in this paper offers the availability model for software aging of SDN controllers by applying software rejuvenation. A stochastic reward net (SRN) is proposed to evaluate the availability assessment of a cluster of SDN controllers. And then how software rejuvenation can improve the performance of SDN controllers is studied. To evaluate the availability of proposed model, mathematical analysis is performed.