OVERVIEW

Networks use switching to achieve connectivity among users while sharing communication links. With switches networks can establish higher-capacity communication paths among users with fewer links and at lower per user cost. In this way networks can take advantage of the economies of scale in communication links.

When we view a switch from the outside as a "black box", it appears as a device with several input and output ports that terminate incoming and outgoing links. An incoming link carriers multiplexed bit streams of several users. The switch guides each of these bit streams to the appropriate output port. Networks differ in the methods used to switch and to multiplex signals. The telephone network uses circuit switching and time-division multiplexing. Data networks use packet switching and statistical multiplexing.

In this chapter we explain the operations and the design of switches, concentrating on two types of switches: circuit switches and packet switches. We will see that large switches are composed of small identical modules. This modular design facilitates development, construction, and testing, and it results in expandable systems.

In section 12.1 we review the tasks performed by switches, and we identify useful measures of performance of switches. In sections 12.2 and 12.3 we discuss circuit switching. The concepts introduced in these sections are also useful in the study of packet switching. In section 12.2 we introduce the principles of time-division and space-division switches. In section 12.3 we present the class of Clos networks, and we explain two key results for these...