Cluster Comput DOI 10.1007/s10586-012-0199-6 Performance implications of non-uniform VCPU-PCPU mapping in virtualization environment

better isolation and consolidation in traditional servers. However, with VMM (virtual machine monitor) layer get-ting involved, virtualization system changes the architecture of traditional software stack, bringing about limitations in resource allocating. The non-uniform VCPU (virtual CPU)-PCPU (physical CPU) mapping, deriving from both the con-figuration or the deployment of virtual machines and the dy-namic runtime feature of applications, causes the different percentage of processor allocation in the same physical ma-chine,and the VCPUs mapped these PCPUs will gain asym-metric performance. The guest OS, however, is agnostic to the non-uniformity. With assumption that all VCPUs have the same performance, it can carry out sub-optimal policies when allocating virtual resource for applications. Likewise, application runtime system can also make the same mis-takes. Our focus in this paper is to understand the perfor-mance implications of the non-uniform VCPU-PCPU map-ping in a virtualization system. Based on real measurements of a virtualization system with state of art multi-core pro-cessors running different commercial and emerging applica-tions, we demonstrate that the presence of the non-uniform mapping has negative impacts on application’s performance predictability. This study aims to provide timely and practi-cal insights on the problem of non-uniform VCPU mapping, when virtual machines being deployed and configured, in emerging cloud