IP網路上傳輸MPEG影像的協定設計研究

Abstract

為了能在IP網路上提供即時性的多媒體應用服務，例如：MPEG影像，需要有適合的即時性傳輸協定和傳輸封包格式。為此，IETF制訂了即時傳輸協定（RTP），RTP封包提供了承載區識別（payload type）和時間戳記（timestamp）。承載區類別指出RTP封包內承載區所承載的多媒體屬於哪一種規格；而時間戳記則提供了這個RTP封包的「時間資訊」。IETF也制訂了即時傳輸控制協定（RTCP）用來監測網路狀況。基於這些IETF RFC文件，我們提出點對點傳輸架構以實現相關的功能。我們討論「TCP友好的」壅塞控制作為傳送速率的依據，並建議未來採用標有優先權的封包，可根據網路狀況選擇丟棄與否，最終希望達到多層可調的傳輸結構。

In order to provide real-time application services (e.g. MPEG video) over IP networks, that needs to define an applicable real-time transport protocol and its packet format. So, IETF define the real-time transport protocol (RTP), and RTP packet header includes the payload type field and the timestamp field. The payload type field indicates which media format this RTP packet carries in its payload. The timestamp field provides the timing information about this RTP packet. IETF also define real-time transport control protocol (RTCP) to monitor the network status. Based on IETF’s RFC specifications, we design a point-to-point transport architecture to implement the associated functions. We discuss the TCP-friendly congestion control mechanism, and determine the sending rate by this mechanism. Finally, we suggest use of the packet with priority in future, and we can choose if we drop the packet according to the network status, so this is a way to achieve layered scalable transporting.