SDN Group Policy Modeling And Network Deployment Testing

With the popularity of the Internet, the users and business have become more complicated. The traditional network architecture exposes various kinds of drawbacks, and it need network technology innovation to ease the pressure. SDN (Software Defined Network) is used in the data center on the characteristics of centralized control, the control and forwarding separation and the open programmable interfaces. In this paper, we are aimed at the status quo and the demand of the power data center virtualization to research on SDN. Combined with the network performance analysis of the power data center, we build a SDN model based on Group Policy. Using the stochastic Petri net model analyses and converts the specific model to get the steady state parameters of each state. Through the concentrated programmable control plane to manage the entire data plane, SDN can satisfy the dynamic characteristics of the virtual environment. The main research contents:1. According to the principle of SDN and virtual technology, aiming at the virtualization application of power data center exist migration problem of IP address conflict and VLAN accessibility limit. Using SDN framework in Data Center Virtualization for modeling, and gives a method of SDN framework that will be applied in a virtual data center.2. By implementing an application centric SDN group policy modeling, divide the physical and logical resources in network. We group the object and policy separately and connect the object through a combination of policy. According to the functions of the OA system to establish a three-layer group policy model include of network, application and database.3. By using performance evaluation algorithm of stochastic Petri net to convert SDN group policy model, and deduce reachability tree-. reachability graph and Markov chain through the process. We can analyze the model steady-state probability, residence time, waiting time and marker density function. Finally we get the system reliability stability to provide a basis for further system optimization.