The Research Of Integrated Multi-Controller Software-Defined Networking Management Platform

With the research of software defined network continues to heat up in recent years, the researchers put a great enthusiasm to this field, especially focusing on the controller. However, the academic and industrial circles haven’t reached a unified standard yet. All want to gain an advantage in the development of controller standard, which results in various types of controllers implemented based on different architectures and programing languages. The complex and not-uniformed control interfaces:commands and functioning methods significantly increase the research difficulty for software defined network. Although most of the controllers are embedded with a visual management platform, it only supports information display. Besides, most of the current researches on software defined network focus on how to improve the performance of the network, how to search for new software defined network applications and so on. But there are few studies focusing on how to simplify the utilization and management of software defined networks.In this paper, we propose an integrated multi-controller software defined network management platform named SDNMP after a deep study of several current controllers. This platform not only supports the visual management of the software defined network, including network topology map creation and display, information retrieve of the network components, addition and deletion of the flow tables, but also unifies the interfaces of the multiple controllers. The first goal is to make it possible for users to focus on controller-related operations based their own needs instead of caring about the type of underlying controllers. This platform is designed as an integration of currently popular controllers but also reserved interfaces for integration of future controllers. Another goal is to make the controller operations easier and more efficient. This platform supports visual interface for user interactions, which greatly saves time for researchers and network administrators comparing to using command-line prompt.Finally, we developed two applications on this platform. One is for network flow statistics, with which users can monitor of network flow changes at any time. Another is to coordinate the management application for multi-controller. This application implements the control power transfer between controllers based on algorithms and methods from foreign researches, which to solve the bottleneck problem caused by single controller. Furthermore, the reserved functional interfaces provide possibility for future platform expansion.