Semi-Markov model-based performance evaluation of a virtualized computing system with aging-induced failures
摘要
In present digital era virtualized computing systems show tremendous potential in tackling computational challenges across diverse real-world systems. However, these systems are susceptible to issues related to aging, which may lead to potential failures. Software rejuvenation is the most preferred approach to address such aging issues. This study proposes two software rejuvenation models of a general virtualized computing system. The system comprises two hosts, with one in active operation while other in backup mode. The hosts are interconnected through a shared memory storage network. The initial model considers repair of Virtual Machine Monitor (VMM) to attains a robust state without rejuvenation after occurance of failure due to aging, whereas the second model requires the VMM rejuvenation following repair after failure, to attain robust state. The theory of semi-Markov regenerative processes is explored model development. The asymptotic performance indices are derived and demonstrated numerically. Evidently, the proposed conceptual models would be essentially important for analysing, designing and implementation of restoration strategies for enhancing reliability and availability of a variety of real-life computing systems.