Research On Edge Caching And Multicast Technology For Wireless Video Services

With the rapid development of online/mobile social media and mobile net-works,diverse mobile video services spring up and lead to mobile video data traffic unprecedented proliferations.It is significant to reduce video traffic re-dundancy by exploiting edge caching and wireless multicasting strategies which can improve the transmission performance of wireless video services and light-ening the burden of wireless networks.The wireless network architectures for typical video services are designed,and the wireless edge caching and sidelink-aided wireless multicast schemes are proposed by utilizing the unique features of video services.The main contributions and novelties of the dissertation are summarized as follows:Firstly,a general wireless video service transmission architecture is de-signed and two proactive edge caching strategies are proposed.By utilizing the powerful computing ability,storage capacity and layered collaboration capabil-ity of Next Generation Radio Access Network(NG-RAN),the wireless video transmission architecture is designed,which includes a cache management con-troller deployed in the Central Units(CU)pool,and cache devices placed in the Distributed Units(DU)and CU pool.Two caching optimization problems of minimizing the average download latency are formulated,and two proactive edge caching algorithms based on joint perception of user mobility and video segment popularity are proposed.Theoretical analyses and simulations prove that,compared with the existing algorithms,the proposed algorithms can re-duce the average transmission latency by about 10%and increase the cache hit rate by about 20%with low complexity.Secondly,an edge caching strategy for Scalable Video Coding(SVC)video service is proposed.A caching optimization problem to minimize the average download latency is formulated,and the user mobility,NG-RAN hierarchical structure and SVC video layered coding are jointly considered in the problem.Aiming at the NP-Hard property,the formulated problem is reconstructed based on the knapsack theory,and a heuristic hierarchical caching algorithm is pro-posed.Furthermore,a simplified algorithm with an approximate ratio of 1/2 optimal solution is proposed for the special scenario,where the remote down-load delay is much higher than the edge download delay.The simulation results verify the effectiveness of the proposed algorithms,and the average download time is 16%lower and the cache hit rate is 30%higher than the existing algo-rithms,respectively.Thirdly,an edge caching strategy oriented to the 360° VR video service is proposed.In view of the latency sensitivity of VR services and the com-putationally intensive feature of Field of View(FoV)synthesis tasks,the FoV synthesis function is deployed in the CU pool of NG-RAN.A hierarchical co-operative caching optimization problem is formulated to minimize the average transmission latency,and a minimum distance maximization caching algorithm based on FoV synthesis is proposed.The approximate ratio of the proposed al-gorithm is theoretically proved.Simulation experiments show that the proposed algorithm can achieve 7%,24%and 60%gain on latency,backhaul traffic and Quality of Experience(QoE)performance,respectively.Finally,a sidelink-aided multicast strategy for multi-quality tiled 360° VR video service is proposed.Considering that the multicast capacity is limited by the users in poor wireless channels,a sidelink-aided multicast transmission sys-tem with two-slot pipeline data transmission scheme is proposed.From the per-spective of system complexity and user fairness,two sidelink-aided multicast scenarios are proposed.A challenging Mixed-Integer Nonlinear Programming(MINLP)problem is formulated,and a two-stage algorithm with greedy search and continuous relaxation is proposed.Theoretical analysis and experimental results show that the algorithms can obtain a sub-optimal solution for multi-quality tiled 360° VR video with low time complexity and fast convergence.