Resource Provisioning in IaaS Clouds; Auto-Scale RAM memory issue

Abstract

In the Infrastructure-as-a-Service model of the cloud computing paradigm, virtual machines are deployed on bare-metal servers called hosts. The host is responsible for the allocation of required resources such as CPU, RAM memory, and network bandwidth for the virtual machine. Thus, the problem of resource allocation reduces to how to place the virtual machines on physical hosts. In this paper, we propose CTMC modeling based on the birth-death process of the queueing systems for the performance of the data center. We will focus on RAM allocation for virtual machines. In this architecture, a job is defined as RAM assignment for a virtual machine. Job arrivals and their service times are assumed to be based on the Poisson process and exponential distribution, respectively. The purpose of this modeling is to keep the number of running hosts minimal in a scalable datacenter while the quality of service in terms of response time is acceptable due to system utilization.