Research On Broadcast Based Performance Optimization For Collaborative Mobile Streaming

In recent years, with the continuous improvement of the performance of mobile devices and the coming of 4G era, mobile Internet started to be widely concerned by academia and industry. As a killer application on the Internet, because streaming media services has high bandwidth, real-time and other characteristics. It has become one of the very challenging mainstream services of the mobile Internet.Traditional mobile stream media services usually use base station access (base-station-to-device, B2D). But this way will cause the network bottleneck with the in-creasing demand of video bandwidth. And in the mobile service environment, because of the constant switching between the base station, the service instantaneous interrupt. And also because of the distance between the user and the base station, the service is not stable. In order to solve the above problems, our project team designed and implemented a system called "Smartphone-Oriented Collaborative Mobile Streaming System". The system connects the mobile phones through Internet WiFi interface, forming a mobile Ad-hoc network. The mobile phones can distribute and share video data in Ad-hoc network. At the same time, the system also supports 4G interface. When the local data can not meet the requirement, it will automatically achieve da-ta from the base station by means of B2D. When the video data being transmitted, the system uses the broadcast monitoring mechanism to reduce the redundant message transmission and reduce the local data transmission overhead.In the "Smartphone-Oriented Collaborative Mobile Streaming System", the re-liability of wireless communication and the design of cooperative cache are two impor-tant factors that affect the performance of wireless communication. The low reliability of the communication mechanism will affect the success rate of data transmission. Effi-cient caching mechanism will make the nodes help each other. For this purpose, in this paper, we study the two problems of performance of wireless broadcast transmission and cache prefetching technology. The main work of this paper is as follows:1) Broadcast monitoring mechanism based on RS coding. In the MAC layer of IEEE802.11, In order to solve the problem of the conflict caused by the hidden termi-nal, the unicast transmission has the RTS/CTS mechanism, but the broadcast transmis-sion doesn’t have the mechanism. So broadcast transmission has low reliability in the wireless environment. Aiming at this problem, this paper proposes a wireless network broadcast transmission scheme based on erasure code. And from the angle of optimiz-ing the network card transmission rate, encoding and decoding and serialization time, the paper gives the optimized implementation mechanism. On this basis, the paper establishes a dynamic encoding optimization mechanism based on feedback combined with the dynamic network link. The experimental results show that the mechanism can effectively improve the overall performance of the radio transmission efficiency and system.2) Data prefetching mechanism based on broadcast monitoring. When users are watching the video, they watch the video for jumping type(VCR) based on their prefer-ences. A large number of VCR operations will increase the user’s waiting delay. Aim-ing at this problem, this paper proposes a data prefetching strategy based on broadcast monitoring. The strategy uses association rule mining algorithm to find the link be-tween user play behavior. On this basis, combined with the characteristics of the wire-less multi hop environment, the strategy effectively prefetch the monitor data blocks. The experimental results show that the mechanism not only improves the overall cache hit rate, reduces the number of communication between nodes and streaming media server, but also effectively saves the disk space of mobile phones and traffic overhead.