The purpose of signalling is to exchange control information between sys-tems of any kind. Familiar applications of signalling include, among others, the control of railway traffic and air traffic. When a pilot asks for permis-sion to land at an airport, he is exchanging signalling information with the control tower. The control tower then provides the airplane with a time slot and a runway on which to land. The presence of signalling in telephony implies that two kinds of traffic are exchanged in a telephone call: sig-nalling traffic—controlling the establishment and release of the voice path
—and voice traffic. We commonly refer to these different kinds of traffic as planes. Thus, two planes can be found in a telephone call: the control plane and the user plane. The control plane handles the procedures controlling the user plane and the user plane handles the actual data or voice trans-mission (see Figure 1-5).
The evolution of the signalling plane has been tightly tied to the evolu-tion of the user plane. New features in the switches could not be exploited without a signalling system capable of taking advantage of them. New sig-nalling systems are designed to make optimal use of the latest advances in the user plane. However, there have also been cases where new signalling systems were introduced without any progress in the user plane. In these situations, some other gain—simplicity, efficiency, robustness—was at stake. To understand how signalling protocols have evolved from the first
telephone system back in the nineteenth century up to the present is to understand which gains were achieved and what exactly triggered protocol design at a certain point of time. SIP is both one such protocol and also a departure from the evolutionary path.
Local and central battery systems After years of using telegraphs, on March 10, 1876, Alexander Graham Bell patented electrical voice trans-mission using a continuous current, and the telephone was born.
Signalling in the early telephone systems was very simple. When the user picked up the receiver, a circuit closed. The terminal then supplied a current to the circuit. The action of closing the circuit meant its seizure. A human operator at the switchboard answered upon seizure of the circuit and routing was accomplished by telling the operator the identity of the callee. The operator switched the call manually. This is the Local Battery system.
Local Battery systems were problematic. Terminals contained batteries, making maintenance more difficult because it was left in the hands of users. Since battery technology was not as advanced as in modern termi-nals, users found the job complicated and were not eager to take it on. The Central Battery system was devised to make things easier on the users.
Central Battery systems supply current to the terminals from the exchange. System automation was the next step in making telephones eas-ier to use. From the installation of the very first exchange in 1878, circuit-switching had been performed manually. In 1889, Almon B. Strowger applied for a patent for an automatic telephone electromechanical switch, and human intervention was no longer needed within the exchanges. Users acquired a dial tone upon seizure of the circuit. After the provision of a dial
Signalling plane
User plane (voice) FIGURE 1-5
Two different planes exist in a telephone call.
tone by the local exchange, terminals forwarded the digits of the callee’s telephone number. With this information in hand, the local exchange routed the call towards the proper destination.
In these first automated systems, the only signalling direction was for-ward, for example, from the caller’s terminal to the switch. Any information that had to be transmitted backwards to the user was sent through the user plane. So if the callee was busy, for instance, a busy tone was sent through the user plane, and it was up to the user to react by hanging up. The control plane never knew if the user hung up because he or she decided to abort the call before the callee answered or because the callee was busy, which proved to be a disadvantage down the line.
DC and AC Analog Systems The next signalling systems to be devel-oped were the DC and AC analog systems, so-named according to the type of current they used. The gain was that DC and AC systems allowed sig-nalling backwards. Having sigsig-nalling in the callee-to-caller direction rep-resents an important saving in some situations, as Figure 1-6 illustrates.
The caller dials a phone number, and each switch in the path between caller and callee reserves a circuit to transport the voice content of this call.
When the signal arrives at the callee’s local exchange, the final switch in the path, it locates the connection to the callee’s terminal. The switch checks the availability of this connection and notices that it is being used for another call at that moment. In a system that does not enable signalling backwards, the local exchange then generates a busy tone. This tone is transmitted to the caller so he or she knows to give up on the call. The caller is expected to hang up after hearing this tone but may not immediately do so. Once he does, signalling indicating the caller has gone is sent forward towards the callee’s local exchange, which at last triggers the release of all the circuits along the path.
It works, but it entails that many circuits in the network remain engaged by this connection even after the exchange reports that the caller is unavailable. DC and AC analogue systems make better use of network resources. Sending busy status through the control plane allows the user plane, circuit-switched path for the call to be released. The busy signal that the caller hears is instead generated by his or her local exchange, and no resources are reserved unnecessarily while the user is informed about the status of the call and prepares to react (see Figure 1-7).
Thus, having signalling in both directions helps manage the network more efficiently. The control plane receives more, and more accurate, infor-mation about the status of the call.