隨著無線網路技術的快速發展，生活周遭中將充斥著越來越多的無線異質網路上。這些無線異質網路所帶來的高頻寬，也使得網路服務蓬勃發展，如網路電話、視訊會議、隨選視訊等多媒體串流服務，但多變的網路環境對多媒體串流的品質的影響將產生下列問題：首先，在進行垂直遞交時，經由不同路徑的多媒體串流在抵達移動式主機同時將造成時間差，若時間差大於影片播放速度，使用者將會察覺到影片跳格的情況。另一方面，因為目前的垂直遞交時間過長，導致移動式主機在進行垂直遞交的同時，使用者將察覺到正在觀看的多媒體串流有跳格或服務中斷等現象發生。 本論文提出一套多媒體串流服務之垂直同步遞交機制。透過機制內的同步預估演算法將可依據目前的網路狀態而預估在不同網路中應向多媒體串流主機所要求的串流畫面順序進而同步經由不同網路的多媒體串流。為了解決因進行垂直遞交所造成的封包遺失，本機制將在進行垂直遞交之前將透過另一張網路接收多媒體串流以減少封包遺失量；另一方面，此時，移動式主機將會同時開啟多個網路介面去接收多媒體串流。因此，將會有重複的封包被接收，本機制也可透過同步預估演算法來有效減少重覆封包的接收量。 本論文將比較多種多媒體串流服務機制與本論文所提的機制做比較，在第四章的模擬中，將可看出所提出的機制在因不同的傳遞延遲而造成的時間差，平均將可改善97.12%。因垂直遞交而造成的封包遺失，本機制只需額外接收約2.1個重複的畫面數，就可平均有效減少72.69%的封包遺失。所收的重複畫面數比起其他機制平均也可減少78.18%。由此可知，本論文所提出的機制將可有效的改善因多變的網路狀態對多媒體串流服務所造成的影響。 With the rapid development of wireless network technologies, there are more and more heterogeneous wireless networks. High bandwidth of wireless heterogeneous networks make possible more and more media streaming services provided in heterogeneous wireless networks, such as the Voice over IP (VoIP), the video conference and the Video on Demand (VoD) and so on. But, media streaming services are fragile and affected easily by varied network qualities. And the following are the major problems to be solved. First, there are very different end-to-end delays in different wireless networks; it causes the offset of frame arrival times when a mobile node roams in heterogeneous wireless networks. Users will be conscious of frame dropped if the offset is too large. Second, the vertical handoff delay is too long and the mobile node may lose many packets in the duration. Users would feel the media streaming services suspended during vertical handoffs. The thesis proposes a Media Streaming Synchronization System (MSSS) which adopts a synchronization estimating algorithm (SEA) to estimate the sequence number of media frame which mobile node should request the media streaming server in different access networks. In other words, the MSSS can synchronize the media streams and then reduce the offset of frame arrival interval. In order to reduce packet loss during vertical handoff, the MSSS receives media streams from another interface before vertical handoff. Because of receiving media streams from multiple interfaces during vertical handoff, there are redundant frames received. The MSSS also reduce the redundant frames during vertical handoff effectively by SEA. In the thesis, the simulations show that the proposed MSSS could reduce 97.12% the offset of frame arrival time. At the part of packet loss during vertical handoffs, the MSSS can reduce approximately 72.69% packet loss and 78.18% redundant frames in average. Hence, the proposed MSSS could reduce the effect of varied network qualities on media streaming services.
