An Analytical Model of IaaS Architecture for Determining Resource Utilization

Abstract

Cloud computing has become a major component of the modern IT ecosystem. A key contributor to this has been the development of Infrastructure as a Service (IaaS) architecture, in which users’ virtual machines (VMs) are run on the service provider’s physical infrastructure, making it possible to become independent of the need to purchase one’s own physical machines (PMs). One of the main aspects to consider when designing such systems is achieving the optimal utilization of individual resources, such as processor, RAM, disk, and available bandwidth. In response to these challenges, the authors developed an analytical model (the ARU method) to determine the average utilization levels of the aforementioned resources. The effectiveness of the proposed analytical model was evaluated by comparing the results obtained by utilizing the model with those obtained by conducting a digital simulation of the operation of a cloud system according to the IaaS paradigm. The results show the effectiveness of the model regardless of the structure of the emerging requests, the variability of the capacity of individual resources, and the number of physical machines in the system. This translates into the applicability of the model in the design process of cloud systems.