Self-adaptive resource allocation for energy-aware virtual machine placement in dynamic computing cloud

Abstract

Cloud computing is a type of Internet-based computing that provides shared computing resources and data on demand. To support the massive cloud services, a data center is built with thousands of servers and switches connected by a communication network called a data center network (DCN). Along with the development of virtualization, the data center manager conducts virtual machine (VM) resizing and live migrations to adjust the allocation of VMs and improve the server utilization. However, compared to VM resizing, live migrations can negatively impact the performance of the system depending on the amount of transferred data and the path length. In this paper, we propose an online resource allocation algorithm using VM consolidation to achieve energy efficiency and reduce service-level agreement (SLA) violations of data centers while considering the power usage of servers, the number of migrations, and the path length of migrations in DCNs. Compared to the state-of-the-art mechanism, the proposed algorithm successfully reduces the power consumption, the number of migrations, and the path length of migrations for a dynamic cloud service.