Title | Service Availability Analysis in a Virtualized System: A Markov Regenerative Model Approach |
Publication Type | Journal Article |
Year of Publication | 2022 |
Authors | J Bai, X Chang, G Ning, Z Zhang, and KS Trivedi |
Journal | IEEE Transactions on Cloud Computing |
Volume | 10 |
Issue | 3 |
Start Page | 2118 |
Pagination | 2118 - 2130 |
Date Published | 01/2022 |
Abstract | With the rapid and wide development and deployment of system virtualization, service availability analysis has become increasingly important in a virtualized system (VS) which suffers from software aging. Software rejuvenation techniques can be applied to improve service availability but its effectiveness depends on the rejuvenation policy, which defines when and where to rejuvenate, and which rejuvenation technique to be triggered. This article aims to analyze the optimal inspection time interval for maximizing application service (AS) availability under a three-level rejuvenation policy, in which rejuvenation techniques are deployed at each level, namely, AS, virtual machine (VM), and virtual machine monitor (VMM) levels. We first apply Markov regenerative process to construct an analytical model for the VS. Experiments of injecting memory leaks are conducted to measure aging-related parameters. Furthermore, numerical analysis is carried out to study the quantitative relationship between AS availability and inspection time interval, and determine the approximate optimal inspection time interval. |
DOI | 10.1109/TCC.2020.3028648 |
Short Title | IEEE Transactions on Cloud Computing |