Live Streaming Delivery Optimization For Multi-Interface Smart Mobile Device

With the rapid development of mobile network technologies and the popularity of smart mobile devices,Various live streaming services based on smart mobile devices become more and more popular.With regard to the data transmmited by live streaming services are mainly video data and the increase of people's demand for live streaming HD video quality,live streaming services always require high-bandwidth support of the network.Although network bandwidth has been greatly improved in recent years,however for a single user,there are still some scenarios with insufficient bandwidth.How to use existing technologies to solve the high-bandwidth requirement of live streaming services remains to be explored.In addition,live streaming services always involve real-time interaction between users.In addition to high video quality,the Quality of user Expericence(QoE)added real-time requirement which needs to be guaranteed by low latency.However,it is a chanllenge to achieve a balance between a high video quality and low latency.Based on the above two challenges of live streaming services,firstly,this thesis intends to aggregate the bandwidth resources of different underlying network interfaces through multi-path transmission technology which based on multi-interface of smart mobile devices to alleviate the high-bandwidth requirement of live streaming services.However,in the mobile scenario,the delay difference of different network interfaces caused a low bandwidth aggregation efficiency of the existing multi-path transmission technology.In this thesis,the data distribution strategy Inidas is designed to optimize the bandwidth aggregation efficiency of current multi-path transmission technology in this scenario.Then,the QoE of live streaming service adds end-to-end latency requirement in addition to high video quality.Considering the conflict between high video quality and low latency,this thesis starts with the broadcast-side of live streaming service architecture,and designs the bitrate adj ustment strategy DGBA to optimize the QoE of live streaming.Finally,the prototype system is implemented according to the optimized strategy.The core part of the prototype system is the data distribution module based on Indias and the bitrate adjustment module based on DGBA.Some proof-of-concept experiments were conducted to verify the feasibility of the two optimization strategies designed in this thesis.By comparing with the current general strategy and commercial strategy,the results show that the two optimizated strategy presented in this thesis do optimize the effeciency of live streaming delivery.