Software Reliability Growth Modelling Framework to Study SDN Controllers

Abstract

Abstract SDN, as an emerging centralized network control platform, has been studied widely by many researchers. To make this complex and fine-grained network platform more reliable and secure, the need to understand its fault debugging pattern was observed. The testing phase of the software has undoubtedly been considered as a crucial phase that ensures high quality network architecture to its users. The pre-release testing activities are solely conducted by the testers where the debugging process after its release comprises both testers and users. The study of the joint role of testers and users has been devoid in the field of network systems. This paper methodically models the complete testing profile (before and after release) of the SDN Controller. The two-staged modeling process is utilized to model the detection and resolution of network faults after its release in which the joint debugging activities of testers and users are catered with the help of Stieltjes convolution. The two-stage model addresses the release point transition. The proposal has been validated on two real life failure data sets. The validation shows promising results.