Research On Segment Scheduling Of SVC-P2P-VoD In Heterogeneous Environment

With the development of the project “broadband China” in our country, thevideo business has gradually become one of the most popular applications. The mobileaccess of smart phones, iPad et al. has become a common application scenario whichbrings new challenges to chunk scheduling of the streaming system. Meanwhile, thesubjective random seeking behaviors of user increase the difficulty for chunkscheduling as well. The main reflection includes the following three aspects. The firstone is that the screen size and its resolution and the processing capability are differentfor different user terminals, and the access bandwidth and Internet traffic charges aredifferent for different access network, which lead to different video quality requirementfor different user terminals. The second one is that the traditional simulcast technologyof multiple coding for heterogeneous environments will causes incensement of burdenon server and obstacle in quality adaptive on the client. The third one is that the randomseeking behaviors of user affect the progress of playback in video, as well as reduce theefficiency of chunk scheduling.However, the existing chunk scheduling technology is mostly appropriate forsingle network environment, it doesn’t combine with SVC segmentation program anddoesn’t take the behaviors of user into account. There is no single integrated programwhich is appropriate for heterogeneous environments and the application scenario ofuser behavior.The researches and contributions of this paper are as follows:①Characteristics of heterogeneous environments, features of P2P network modeland P2P system, SVC scalable video coding techniques are studied. Based on the aboveanalyses, the problem of different video quality requirements in heterogeneousenvironment is solved. Since the traditional program which multiple coding for thesame video has low scalability, the best matching program adopted in P2P streamingsystem is SVC program with one time scalable coding.②The data transmission architecture of video data in SVC-P2P streaming systemis studied. Then the existing chunk segmentation strategy and corresponding chunkscheduling strategy are analyzed. By considering the characteristic of SVC layers andthe random seeking behaviors of user, the Anchor-P2PSVC Chunk SchedulingAlgorithm based on User Behavior Characteristic (UACS algorithm) is designed. In the proposed algorithm, two types of windows are designed. One is playback window basedon playback time slot for ensuring continuous data of ordinal playback, the other one isanchor window designed with data prefetching, which based on user random seekingbehaviors following the Weibull distribution. Then, the layer-by-layer schedulingstrategy is utilized in playback window and the first anchor window to ensure thetimeliness of data, and the rarest-first strategy is used in the other anchor windows tobalance the chunks distribution of the whole system.③The chunk scheduling module on client in SVC-P2P video on demand streamingsystem (SVC-P2P-VoD) is improved. Moreover, the simulation platform based on thissystem is set up and the simulation of the heterogeneous environment of three types ofterminals (mobile phones, laptops, PC) with three types of networks (wired, WLAN,3G) is implemented. Furthermore, comparison of the UCAS algorithm and LLalgorithm and SCS algorithm in the application scenario with the user random seekingbehaviors is completed. Result shows that the proposed algorithm can improve thechunk scheduling performance, shorten the response time, and enhance the quality andfluency of users in watching video.