Cloud Computing provides infrastructure as a service (IaaS) for users based on virtualization technology. The load of applications and services on IaaS platform is dynamic, thus its demand of virtual resource is also dynamic. Therefore, to collect and analyze the virtual resources footprint of in the platform, and flexibly scheduling them on demand has become the key point to improve overall service performance and resource utilization ratio of the cloud computing platform. This paper, with the concern of load balancing and reducing the user cost of the cloud platform, brings forward an elastic cloud structure based on the load of virtual machine in the platform. Simulation results show that the method can improve the utilization ratio of virtual resources and reduce user cost.