Distribution of Virtual Machines in Cloud Infrastructure with the Account of Various Posibilities of Access to Data Storage Systems

Abstract

the article addresses the well-known "knapsack problem" in the context of the distribution of virtual servers within the cloud infrastructure. At the same time, the main focus in solving this problem is the distribution of virtual machines within the cloud, which includes resources with different access capabilities. The mathematical model proposed in the article allows conducting optimal allocation of resources not only taking into account such parameters as random access memory and processor power, but also parameters of data storage systems related to memory capacity and access speed. The resulting model was tested on a real network infrastructure. From a practical perspective, the method of virtual machines distribution under question allows using the low-power devices available at the data center.