Modeling and Evaluation of Software Rejuvenation in Cloud Computing

Please enable javascript in your browser.

Fakhrolmobasheri, Sharifeh | 2017

526 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 50257 (19)
University: Sharif University of Technology
Department: Computer Engineering
Advisor(s): Movaghar, Ali
Abstract:
Long and continuous running of software is likely to cause software aging-induced errors and failures. Cloud service centers suffer from such errors because of degradation of virtual machine monitors (VMMs) that oversee the execution of virtual machines (VMs). The proposed solution to prevent such VMM failures is to use a rejuvenation scheme for planned termination and restarting of monitor software. The best measure to determine the optimum time of rejuvenation is the VMM’s age, that is, the cumulative workload carried out since the last reboot, repair, or rejuvenation. This paper provides an analytical model based on stochastic activity network (SAN) tool for performance evaluation of cloud service centers based on VMMs’ age. The modeling results show that implementation of rejuvenation scheme improves the data center’s availability and reduces the average rate of failure in VMMs
Keywords:
Performance Modeling ; Virtual Machine ; Cloud Computing ; Performance Evaluation ; Software Rejuvenation ; Virtual Machine Monitor

No TOC