Research On Cooperative Video Caching And Delivery Schemes In Mobile Edge Networks

With mobile streaming services increasingly popular and mobile users preference for high quality videos,the traffic of video data on wireless networks is exponentially in-creasing.In the traditional centralized service architecture,each mobile user requests video from the streaming server independently via cellular base station,resulting in enormous bandwidth overhead faced by streaming service providers.Meanwhile,re-dundant transmissions of popular videos overwhelm the throughput that cellular base stations could afford,and cause congestions on back-haul links between streaming servers and base stations,which will seriously degrade the quality of streaming ser-vice.In mobile edge networks,one promising solution to cope with this problem is cooperation between edge mobile devices and edge network terminals,which can of-fload the caching and delivery of video data to locations closer to streaming users,such as D2D(device to device)networks and small base station networks.Mobile devices or small base stations download and cache video data in a cooperative manner,and delivery video data to streaming users through local links,which can not only reduce the delay of video transmission,but also can eliminate the redundant transmissions of popular videos over back-haul links,and finally improve the quality of the streaming service and the scalability of the streaming system.In this dissertation,we conduct a comprehensive study on the technologies of cooperative video caching and delivery in mobile edge networks,with the traffic offloading of streaming servers as the main motivation,and with the scope and object of cooperation as the researching clue.For the following three network scenarios,i.e.,synchronous video(users with the same playback offset)in one-hop D2D network,asynchronous video(users with different playback offsets)in multi-hop D2D network,and multiple bitrate video in small base station network,we derive the bottleneck resources as wireless channel,available ener-gy for video transmission and storage capacity,respectively.Then,we investigate the cooperative caching and delivery schemes over these network architectures.We aim to explore and understand the cooperation mechanisms of video caching and delivery under different network architectures,and establish the relationship between multi-dimensional resources and the quality of streaming service.The contributions of this dissertation are summarized as the following:(1)We first consider a scenario where a group of mobile devices are close to each other and are interested in the same video.Mobile user can communicate to each other via D2D connections,so that the video can be downloaded in a cooperative fashion.In one-hop D2D network,traditional video delivery schemes usually consider a single D2D communication channel.In order to make better use of the communication re-sources in the D2D network and explore the transmission capabilities of multiple D2D link interfaces,we investigate the video delivery scheme based on the allocation of multiple D2D channels.Due to the small size of the network,we assume that the time duration of download stage is slotted,and the transmissions over D2D communications are scheduled in a centralized fashion so as to achieve a high transmission efficiency.We investigate the allocation scheme of multiple channels on each time slot.Aiming to minimize the content download time,we formulate this problem as an integer linear programming problem and show its hardness and inapproximability.For the case that only a single channel is available for D2D communications,we propose an asymptot-ically optimal algorithm.For the multi-channel case,we further propose a heuristic algorithm with low time complexity and demonstrate its efficiency via extensive simu-lations.(2)Further,we extend the cooperative delivery scheme to a larger multi-hop D2D network where mobile users with heterogeneous data requests,i.e.,users have differ-ent playback offsets of the same video.Unlike the one-hop D2D network,the delivery of video in a multi-hop network usually needs the assistances of multiple relay nodes,and the video data transmission over multi-hop D2D links would introduce addition-al energy consumption on involved intermediate mobile devices.Under the energy constraint of individual mobile device,we investigate delivery schemes to minimize the overall traffic of the cellular network by efficiently disseminating video over the multi-hop D2D network.We formulate this video cooperative delivery problem as an integer linear programming problem,and show that it is NP-hard and even hard to approximate.Then,we present an energy allocation based algorithm and a simulated annealing heuristic algorithm,and demonstrate that there exist a trade-off between the performance and complexity of this two approaches.We also evaluate the performance of our proposals via both theoretical analysis and extensive simulations.(3)Finally,we focus on a streaming system where numerous videos are served and conduct a comprehensive study on cooperative caching and delivery scheme for videos with multiple bitrate supported in small base station networks.In this scenario,each small base station is attached by a mobile edge server,which is capable of high rate transmissions to nearby mobile users,but with limited storage capacity,and our ob-jective is to maximize the mobile user perceived QoE with the assistance of small base station caching.With the multiple bitrate encoding techniques supported,caching de-ployment is challenging for the following two facts.First,cooperative caching should be explored for mobie users located at overlapped coverage areas of small base station-s.Second,there exists trade-off consideration for caching either high bitrate videos or high diversity videos.By formulating the QoE value as a function of video bitrate received by mobile user,we derive the relationship among video bitrate selection,user perceived QoE and caching scheme,and then propose several approaches for coopera-tive cache deployment schemes.Aiming to maximize the mobile user perceived QoE,we formulate the multi-bitrate video caching problem,and prove this problem is NP-hard for any given positive and strictly increasing QoE function.We then propose a polynomial complexity algorithm,which can achieve an approximate ratio arbitrari-ly close to 1/2 for any positive and strictly increasing QoE function.Specifically,for a linear QoE function,we explore special properties of optimal solutions,based on which more efficient algorithms are proposed.We demonstrate the effectiveness of our solutions via both theoretical analysis and extensive simulations.