317Applications provided by service providers have some history, and it is
remarkable that even in the IP environment, many wireline and wireless ser-
vice providers have not yet learned from this history. This is the reason why, in
this chapter, we will show how value-added services by providers using SIP
can be implemented, using application servers in an open, distributed, and
loosely coupled architecture that is highly scalable. The application server
approach is based on the client-server (CS) model for SIP.
Using application servers in the network is, however, not the only approach.
In Chapter 20, “Peer-to-Peer SIP,” we will show how applications can also be
implemented in peer nodes. In the extreme, the most frequent applications can
reside entirely in the endpoints, or, in a mixed environment, some applications
can also reside in P2P SIP supernodes, where the architecture described in this
chapter will apply.
NOTE On a historical note, among issues not to forget is that value-added
services in the PSTN are implemented using the Intelligent Network (IN) [1]
based on central control. The IN is a collection of servers and other resources
used to control call setup and to provide voice features, such as announcements,
voicemail, and so on. In hindsight, IN services seem rather frugal compared with
communications on the Internet. Other architectures, such as H.323 or the so-
called “softswitches” based on IP telephony gateway decomposition [2], have
similar approaches to the IN for enhanced voice features. We use the term