In this chapter, we set out the main issues of queueing theory (or the theory of queueing networks). Stemming from probability theory, as with reliability and statistics already presented, its application to the field of telecommunications for the solving of traffic problems has given rise to a well-known discipline: teletraffic.

Queueing theory sets out to take account of the phenomena of waiting and congestion that are linked to the random and unforeseeable nature of the events encountered (in the case with which we are concerned, connection setup requests, IP packets going through a router, tasks to be processed by the processor of a server, etc.). In view of the problems to be resolved as set out in the chapter on quality of service (waiting times to be complied with, loss probability in buffers, etc.), the importance of this theory will easily be understood from the viewpoint of our essential aim: to assess the performances of systems and networks.

Whenever a theory is proposed, it offers a schematic and simplified vision of observed reality. In point of fact, this theory will be used on models, abstractions from the actual world which capture the essence of the phenomenon to be studied. The activity of modelling, which leads from the actual system to the queueing model, takes on essential importance in this case.

Obviously, the theory does not take into account all the diversity of reality - or more accurately is not able to respond to it with perfect fidelity.