The Study Of The Models And Algorithms Optimization For The I/O Scheduling And Smoothing Transmission Of The VBR Streaming Dideos

With the increasing of network bandwidth, applications of network streaming media have been mainstreams on network. Technique relations to network streaming media have been studied by many researchers in computer science. Videos are main data source of network streaming media, For example VoD, network videoconference, and tele-learning need real transmission and transmitting of videos. Compared to the CBR compressed video, the VBR compressed video can achieve more satisfying QoS under the same resource constraints. The disadvantage is that the VBR characteristic of the VBR compressed video complexes network management techniques. This dissertation focuses on two main problems: the I/O scheduling and smoothing transmission of the VBR video stream.The dissertation first addresses two problems: one is the maximum output of video streams with multi-network I/O channels, and proposes a real streams scheduling model with buffer box, and develops an admission control algorithm base on this model. Another is the study of the greatest possible reduction to VBR video when transmitting stored video across a network to a client with given buffer size, and develops a shortest path smoothing algorithm. For these two algorithms, we present formal analysis of optimality, and experimental results show that the algorithms are efficient.Base on the shortest path smoothing algorithm, the dissertation proposes an asynchronism smooth scheduling, which considers the heterogeneity transport characteristic between two networks, and a transmission service protocol, which can provide real-time video service under the residual bandwidth resource, and a real-time smoothing algorithm, which considers the delay character of the strong real-time video. Experimental results show that these algorithms work very well.After discussing a video scaling placement algorithm, the dissertation presents some disk retrieval algorithms and video divided schemes aiming to optimize performance of the disk I/O, and to integrate the following techniques: smoothing, video placement and disk multi-rates. The simulation results show that these algorithms are efficient.