Poster: A Software-Defined Multi-Camera Network

Abstract

The widespread popularity of OpenFlow leads to a significant increase in the number of applications developed in SoftwareDefined Networking (SDN). In this work, we propose the architecture of a Software-Defined Multi-Camera Network consisting of small, flexible, economic, and programmable cameras which combine the functions of the processor, switch, and camera. A Software-Defined Multi-Camera Network can effectively reduce the overall network bandwidth and reduce a large amount of the Capex and Opex when we set up the multicamera system. Furthermore, we propose a centralized management system along with a Web User Interface for the user to view the information from the GPS sensor and the digital compass, and monitor the range through the Google map. The user also can see the real-time video stream easily via clicking the camera icon. We propose managing a large amount of cameras, owned by different enterprise customers, through the authentication mechanism and the function of network virtualization in Software-Defined Networking. The groups of customers from different enterprises can manage their cameras in their own Virtual Multi-Camera Networks, isolated from the other enterprise customer’s network. Therefore, we can provide different qualities of service through the network virtualization to different enterprise customers. In addition, the forwarding path of the video stream can be dynamically changed by the SDN controller to achieve a seamless real-time video stream, and support an efficient multicast of the video stream and user mobility. Our goal is to address the problems which cannot be programmed, and that have been found difficult to solve in the legacy network over the years through our Software-Defined Multi-Camera Network platform. We designed a smart campus camera network application for real time video surveillance service on the top of the SDN controller, the video seen by school security in the Web User Interface will be dynamically updated if the tracked student moves from the monitoring range of one camera to another.