Semi-Markov's models for availability assessment of an Infrastructure as a Service Cloud with multiple pools of physical and virtual machines

摘要

It is well known, that different types of physical (PMs) and virtual (VMs) machines failures and as well deliberate malicious impacts on information resources of cloud infrastructure are serious obstacles to development of whole cloud ecosystem. What is more, in the current environment of fast growing number of cloud subsystems and supporting devices, the cloud service providers need to analyze and control overall availability level of cloud infrastructure. In this paper we summarized the basic provisions of approach to assess availability of an Infrastructure as a Service (IaaS) Cloud that is implemented based on Semi-Markov modeling process (SMP). This paper is our real attempt to improve the modeling process of availability for IaaS Cloud considering different negative impacts on resources of PMs and VMs, namely sudden failures, hacker's attacks, operator's and programmer's mistakes.