Int J Performability Eng ›› 2018, Vol. 14 ›› Issue (9): 2083-2089.doi: 10.23940/ijpe.18.09.p17.20832089

Previous Articles     Next Articles

Performability Modeling for Cloud Service with Check-Pointing Mechanism Considering Hardware and Software Failures

Xiwei Qiu, Liang Luo*, Sa Meng, and Xiaochuan Tang   

  1. University of Electronic Science and Technology of China, Chengdu, 611731, China
  • Revised on ; Accepted on
  • Contact: * E-mail address: 376454384@qq.com

Abstract: Cloud service performance is an important metric that must be considered in detail. Most existing researches study various methods and approaches for evaluating the performance metric; however, these are inadequate because they do not take into account dynamic performance changes caused by reliability factors. In fact, both software failures of a virtual machine (VM) and hardware failures of a server inevitably interrupt the execution of a cloud service and eventually result in more time being spent on completing the cloud service. Meanwhile, the check-pointing mechanism is an important fault tolerant technique that is widely adopted to handle software failures. In this paper, we present a joint modeling approach encompassing Semi-Markov and the Laplace-Stieltjes transform to analyze the reliability-performance correlation for cloud services that adopt the check-pointing fault recovery mechanism. Finally, we present a recursive method to evaluate the expected service time.

Key words: cloud service, check-pointing mechanism, hardware failures, software failures