Research On Streaming Transmission Optimization For Mobile Devices

With the rapid development of the cellular technologies,the streaming application on mobile devices gains increasing popularity these days,whose traffic has accounted for more than 55%of the total network traffic.Streaming is a kind of application that the data can be used while downloading.Due to the video length,streaming needs to occupy the transmission channel for a long time.However,because of the unreliable data transmission in cellular networks and the energy limitation in the mobile devices,mobile streaming cannot provide the satisfying playback quality to users.In this paper,we try to schedule the streaming data transmission to improve the service of streaming application.And the paper mainly focuses on three issues(streaming rebuffering time reducing,transmission energy consumption saving and DASH bitrate selection).Firstly,due to the changeable throughput of the mobile devices,it is hard to guarantee the smooth playback.Based on the rebuffering time model design,we propose an online scheduling transmission strategy RTMA to improve the playback fluency in multi-users scenario,which can be proved to be local,optimized.The simulations show that RTMA can reduce 68%average rebuffering time than the existing strategies.Secondly,the energy consumption is the bottleneck of the development of mobile applications.According the measurement research,we find that streaming spends more than half of the energy consumption on the data transmission.In order to improve the energy efficiency while ensuring the smooth playback,we first model the energy consumption optimization problem in multi-user scenario and creatively raise a rebuffering-time-queue based online optimization solution EMA.This paper also evaluates the performance of EMA from both the mathematic analysis and simulation.Based on the simulation experiments,we find that EMA can reduce energy consumption greatly compared with SALSA,EStreamer and Default.In the last,DASH applications are aimed to change the streaming bitrate according to the devices' throughput.However according to the research of the present bitrate selection mechanisms,we find that these mechanisms cannot choose the most suitable bitrate for users,which will result in frequent bitrate switch or low video bitrate.Using the data remaining in the buffer,we design a dynamic bitrate selection mechanism DBS to reduce the bitrate switch frequency and enhance the quality of users' experience in DASH.The simulations shows that DBS can adapt to the variable channel condition very well and also improve the video playback quality.