Optimizing software rejuvenation policy for real time tasks

• Published in: Reliability engineering & system safety Vol. 176; pp. 202 - 208
• Main Authors: Levitin, Gregory, Xing, Liudong, Ben-Haim, Hanoch
• Format: Journal Article
• Language: English
• Published: Barking Elsevier Ltd 01.08.2018; Elsevier BV
• Subjects: Aging; Computer programs; Downtime; Event transition; Failure analysis; Optimization; Performance degradation; Product testing; Real time; Real time task; Reliability engineering; Software; Software engineering; Software rejuvenation; System effectiveness; Systems analysis; Time dependence; pdf; PS; cdf; RDF; Software rejuvenation; Real time task; Event transition; GA; ETL; Optimization; PTC
• ISSN: 0951-8320; 1879-0836
• DOI: 10.1016/j.ress.2018.04.010

•Rejuvenation in software systems with multiple degradation levels is considered.•New method for evaluating success probability of real time tasks is suggested.•Arbitrary time-to-state-transition distributions are allowed.•State-based rejuvenation policy optimization problem is formulated and solved. Software rejuvenation is a proactive maintenance technique adopted in diverse and wide applications for mitigating performance deterioration effects of software aging and further preventing the system crash from happening. As the software rejuvenation procedure incurs system overhead and downtime, it is relevant and crucial to optimize the software rejuvenation policy to maximize its benefit and effectiveness. This work considers the optimal rejuvenation policy problem for systems subject to multiple performance degradation levels and performing real-time tasks. The solution encompasses a new iterative method for time-dependent evaluation of the task successful completion probability for the considered real-time systems, enriching and expanding the existing evaluation methodologies for software aging and rejuvenation systems. Based on event transitions, the proposed evaluation method is efficient and applicable to arbitrary types of state transition time distributions. The state-based rejuvenation policy is further optimized to maximize the probability of the task completion by the predetermined deadline. Examples are provided to illustrate applications of the proposed methodology as well as effects of different system parameters on the optimization solution.