Implementation for Video Streaming Control Prototype System Based on OpenFlow
|School||University of Electronic Science and Technology|
|Course||Applied Computer Technology|
|Keywords||video streams interest mining OpenFlow OPNET|
Related data in Chinese Internet platform shows that the network video is the application whose users visit time per day is the longest, meanwhile online video has become the largest Internet application. For network operators, its main revenue comes from the service time for the users. While a mass of video flow has not brought the corresponding revenue growth, at the same time the whole industry tariff level falls fast due to the intensifying market competition. Facing with the higher operation and maintenance expenses, and the lower profit margins, operators need to change the traditional profit model, whose key is the shift from business to service. For video service providers, advertising revenue is most important for their development as the current video offering is free. And the major and stable video user groups is the guarantee of the revenue. Obviously, the attractive information is critical in cultivating loyal fans especially when video resources are exceptionally rich. So if we can accurately initiative recommended video based on user interest, it is strongly helpful for us to capture video users and improve the profit. However, if the network conditions affect the ultimate visual experience, what we did will fall short.On the other hand, if the operator’s network equipment can be aware of the business, it is possible to discover more users’ interest through the video streams in the network. So differentiated services and charging problems can solved by providing personalized services. Video service providers can also be customized a higher quality of network service and provide a better user experience, so the value‐added services provided by operators and value‐added costs paid by video service providers, will be a win‐win situation.This thesis proposed to establish a video stream based on OpenFlow control system in the network aggregation layer environment, which is based on in‐depth analysis on current video stream recognition and interest mining techniques. Main work of this thesis is:1. Research on video streams recognition and interest mining techniques, analyze the OpenFlow technology architecture theory and composition based on characteristics of this control system, and give the feasibility of the prototype system under OpenFlow framework.2. Design a system, its main function is to recognize P2P, Web video streaming using related rules and algorithms when the network video stream transmits via convergence devices, routers. on this basis, the system aims to extract the video interest information according to the resulting information and user interest model, as well as conduct priority scheduling to the identified video streams, then design an appropriate billing function.3. Build a test environment on the OPNET network simulation platform, and test and analyze the system.