A self-aware strategy for virtual machines placement on clouds

Fen Guo¹, Huaqing Min¹, Ming Yin²

¹School of Software Engineering, South China University of Technology, Guangzhou Higher Education mega center,510006, Guangzhou, China
²School of Automation, Guangdong University of technology, Guangzhou Higher Education mega center,510006, Guangzhou, China

Cloud computing is a new computing service mode, and virtualization is a key technology of it. A self-aware strategy (SAST) for Virtual machines (VMs) management on clouds is proposed which is multi-attributed weighted on the resources. It manages the virtual resource basing on the requests of users and the real-time state of the system dynamically. It consists of three phases: (1) monitoring the cloud performance including VMs and Physical Machines (PMs), with the data standardized; (2) measuring the cloud load balance value with the attribute weighted measurement model; (3) using the placement algorithm to choose the best appropriate PM to place the VM requested. The main contribution of the paper is that a cloud load balance measurement model is introduced and a VM scheduling strategy is proposed which includes the VMs placement optimization algorithm and the VMs dynamic migration algorithm. The SAST is tested on the simulation platform comparing with other traditional ones. As a result, we concluded that it guaranteed the SLA and achieved better load balance of cloud. And at the same time, it minimized the number of the started PMs on clouds to reduce energy consumption.