Overview

The foregoing chapters address the issues related to Internet Protocol (IP) networks in general and the mechanisms for transporting digitally encoded voice across those networks. Chapter 2, "Transport Voice by Using IP," in particular, describes how voice is carried in Real-Time Transport Protocol (RTP) packets between session participants. What is not addressed, however, is the setup and teardown of those voice sessions. Thus far, we have assumed that session participants know of each other's existence and that media sessions are somehow created such that they can exchange voice using RTP packets. So how are those sessions created and ended? How does one party indicate to another a desire to set up a call, and how does the second party indicate a willingness to accept the call? The answer is signaling.

In traditional telephony networks, specific signaling protocols are invoked before and during a call in order to communicate a desire to set up a call, to monitor call progress, and to gracefully bring a call to an end. Perhaps the best example is the ISDN User Part (ISUP), a component of the Signaling System 7 (SS7) signaling suite. In Voice over IP (VoIP) systems, signaling protocols are necessary for exactly the same reasons.

The very first VoIP systems used proprietary signaling protocols. The immediate drawback was that two users could communicate only if they both used systems from the same vendor. This lack of interoperability between systems from different vendors was a major inconvenience...