The All-New Switch Book: The Complete Guide to LAN Switching Technology, Second Edition

1.2. Addressing

By definition, a network comprises multiple stations.^[] The purpose of the network is to allow information exchange among these multiple stations. An address is the means used to uniquely identify each station either as a sender or receiver of information (or both).

^[] A network consisting of one station is about as useful as a single walkie-talkie.

Every layer that supports data exchange among multiple stations must provide a means of unique identification, that is, some form of addressing.^[] Many Data Link technologies (for example, LANs) allow multiple devices to share a single communications link; Data Link addresses allow unique identification of stations on that link. At the Network layer, you need to uniquely identify every station in a collection of multiple, interconnected links. Therefore, most network architectures provide for station addresses at both the Data Link and Network layers.

^[] Strictly speaking, station addresses are necessary only when communications can occur among some proper subset of the totality of stations present. For example, if every transmission is always intended for receipt by every station, then there is no need to identify the target receiver(s). However, this is a rather unique case, somewhat artificially contrived. In practice, we need to provide station addresses at every layer where there are multiple communicating stations.

1.2.1. Local and Global Uniqueness

The only important characteristic of an address is its uniqueness; ...