網際網路語音協定界接與附加服務

Abstract

隨著網際網路的普及，以網際網路為平台傳送語音已愈受歡迎。網際網路電話擁有傳統電話難以企及的優勢，諸如成本低廉，易於開發新服務，以及提供語音與數據整合應用。目前主要的VoIP協定有International Telecommunication Union (ITU) 提出的H.323，和Internet Engineering Task Force (IETF)提出的Session Initiation Protocol (SIP)、Media Gateway Control Protocol (MGCP)、以及MEGACO，這些方案各具優劣，但整合這些方案卻為當務之急。 我們提出一個界接架構，使得不同的協定間可彼此溝通。界接功能是利用有限自動機來完成不同協定間的訊息轉換，我們將此功能實作於call agent中。為了減低call agent的負擔，我們選擇SIP作為call agent間溝通的媒介，以達到分散處理的目的。我們也提供了新增服務的機制，使得新服務的開發與實作更添效率。最後，我們也提供實作，以驗證整體架構的可行性。

As the widespread of the Internet, utilizing the Internet as a bearer to transport voice traffic becomes more and more popular. Internet telephony offers a number of advantages over the traditional circuit-switching network, such as lower cost, easy to implement services, integration of voice and data applications. There are several Voice over Internet Protocol (VoIP) standards proposed, including H.323 proposed by International Telecommunication Union (ITU), Session Initiation Protocol (SIP) proposed by Internet Engineering Task Force (IETF), Media Gateway Control Protocol (MGCP), and MEGACO proposed by IETF and ITU. These approaches have their own advantages and limitations. It is critical to provide a mechanism to enable communication between different protocols and networks. In this thesis, we propose an interworking architecture. The interworking function is a collection of finite state machines in a call agent, and it can convert the signaling of different protocols. To balance the load of call agents, SIP is used as the signaling protocol between call agents. We also propose a mechanism to provide supplementary services. We implement the design on a VoIP platform to demonstrate the feasibility of the architecture.