Font Size: a A A

Research On Trsansmittion Control Of Multi-party Interactive Video Streaming

Posted on:2022-01-23Degree:MasterType:Thesis
Country:ChinaCandidate:X Y HuFull Text:PDF
GTID:2518306341951629Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Multi-party interactive live streaming is a new commercial live streaming mode,which is more and more worldwide popular,and has promising prospects and value.However,due to the origin of multi-party interactive live streaming from the traditional single flow live streaming technology,there is a large room for improvement in service performance and user experience.At present,the transmission control of multi-party interactive live streaming is mostly the reuse or simple integration of traditional live streaming scenes.How to further optimize the control methods and strategies is of great significance to improve user experience.As an early exploration of multi-party interactive live streaming transmission control,this paper studies the optimization of user experience quality prediction model and the frame scheduling strategy in the forwarding servers and its system architecture.This paper mainly includes following contributions:(1)In order to better improve the quality of user experience,this paper proposes a prediction QoE model based on video content,The model optimizes the original model referring to the latest research on QoE,so as to improve the representation ability of the model and make the prediction of the model closer to the users' perceptive experience.(2)This paper proposes a scheduling algorithm,Rapunzel,for multi-party interactive live streaming,which is deployed on the forwarding server.Rapunzel predicts the QoS(Quality of Service)terms of different scheduling strategies based on queuing theory model,and combines with the QoE model to obtain the optimal scheduling strategy.Finally,this paper verifies the proposed QoE prediction model optimization scheme through numerical experiments and measurements,and tests the performance of the Multi-party interactive video streaming scheduling algorithm Rapunzel in real network data through a simulation system.Experiments show that the optimized QoE prediction model has a stronger ability to characterize the quality of user experience,and the scheduling algorithm Rapunzel can effectively solve the technical bottlenecks in multi-party interactive live streaming,effectively reduce the probability of stall,and improve overall users' experience...
Keywords/Search Tags:streaming, quality of experience, scheduling schemes, queueing theory
PDF Full Text Request
Related items