The Research And Design Of Publish/Subscribe Management System Based On Openflow

With the rapid growth of network scale and application variety, the structure and function of traditional network become increasingly complicated, which makes it hard to manage. SDN(Software Defined Network) is a new type of network architecture, and OpenFlow is an industry-standard SDN communications protocol. It separates the control function from network devices, which realizes flexible control over data packets of network and provides a new platform for network applications. This paper focuses on the data timeliness and security of the Publish/Subscribe system, uses OpenFlow switch simulated by OVS and Floodlight controller to build a new network environment realizing centralized management. The system is used for managing topological information, device information, network flow information, subject tree information, user subscription information and strategy information.Meanwhile, the Publish/Subscribe Systems often suffer from subscription attack. As for security, this paper analyzes user feature, finds suspicious users using data mining algorithm and impose restrictions on them to protect the system from subscription attacks. The data mining algorithm used in this paper is K-Means algorithm which is a clustering algorithm. Finally, we build a more manageable, secure and reliable Publish/Subscribe System. At the end, we make experiments to test the Publish/Subscribe Management System and the Defense Mechanism against subscription attacks. The results show the mentioned features have been realized.