Abstract:
The loss of a node in a computing network, either for unplanned reasons, such as hardware failure, or planned outages, such as upgrades, can result in degraded computing network performance or loss of redundancy. To reduce this possibility, a spare node is kept powered on and visible on the network. In this paper, we study the Markovian warm-spare nodes provisioning computing network where spare nodes may also be predisposed to failure in switching from standby state to the operating state. In addition to switching failure and common cause failure, the realistic and economical maintenance server’s modified multiple working vacation policy and failed nodes’ synchronized reneging are also considered. Reliability characteristics of the computing network for I/O operations have been derived using transient-state probabilities which have been computed using the theory of the Quasi-Birth-and-Death process, Laplace transforms, Eigenvalue and Eigenvector. The critical analysis of reliability characteristics has also been done and the paper has been enriched with numerical results in the form of tables and graphs to provide a glance at the investigation. The concluding remarks and future scope have also been included.