A computer network is the infrastructure that allows two or more computers (called hosts) to communicate with each other. The network achieves this by providing a set of rules for communication, called protocols, which should be observed by all participating hosts. The need for a protocol should be obvious: it allows different computers from different vendors and with different operating characteristics to ‘speak the same language’.
1.1 Abstract Network
Figure 1.1 shows an abstract view of a network and its hosts. The network is made up of two types of components: nodes and communication lines. The nodes typically handle the network protocols and provide switching capabilities. A node is usually itself a computer (general or special) which runs specific network software.The communication lines may take many different shapes and forms, even in the same network. Examples include: copper wire cables, optical fiber, radio channels, and telephone lines.
A host is connected to the network by a separate communication line which connects it to one of the nodes. In most cases, more than one host may be connected to the same node. From a host’s point of view, the entire network may be viewed as a black box, to which many other hosts are connected. Each host has a unique address allocated to it by the network. For a host to communicate with another host, it needs to know the latter’s address. All communication between hosts passes through the nodes, which in turn determine how to route messages across the network, from one point to another.
Figure 1.1 An abstract network.
1.2. Network Types
Networks may be divided into different types and categories according to four
1. Geographic spread of nodes and hosts. When the physical distance between the hosts is within a few kilometers, the network is said to be a Local Area Network (LAN). LANs are typically used to connect a set of hosts within the same building (e.g., an office environment) or a set of closely-located buildings (e.g., a university campus). For larger distances, the network is said to be a Metropolitan Area Network (MAN) or a Wide Area Network (WAN). MANs cover distances of up to a few hundred kilometers and are used for inteconnecting hosts spread across a city. WANs are used to connect hosts spread across a country, a continent, or the globe. LANs, MANs, and WANs usually coexist: closely-located hosts are connected by LANs which can access hosts in other remote LANs via MANs and WANs, as illustrated in Figure 1.2.
2. Access restrictions. Most networks are for the private use of the organizations to which they belong; these are called private networks. Networks maintained by banks, insurance companies, airlines, hospitals, and most other businesses are of this nature. Public networks, on the other hand, are generally accessible to the average user, but may require registration and payment of connection fees. Internet is the most-widely known example of a public network. Technically, both private and public networks may be of LAN, MAN, or WAN type, although public networks, by their size and nature, tend to WANs.
Figure 1.2 Example of a WAN between LANs
3. Communication model employed by the nodes. The communication between the nodes is either based on a point-to-point model or a broadcast model (see Figure 1.3). In the point-to-point model, a message follows a specific route across the network in order to get from one node to another. In the broadcast model, on the other hand, all nodes share the same communication medium and, as a result, a message transmitted by any node can be received by all other nodes. A part of the message (an address) indicates for which node the message is intended. All nodes look at this address and ignore the message if it does not match their own address.
Figure 1.3 Communication models.
4. Switching model employed by the nodes. In the point-to-point model, nodes either employ circuit switching or packet switching. Suppose that a host A wishes to communicate with another host B. In circuit switching, a dedicated communication path is allocated between A and B, via a set of intermediate nodes. The data is sent along the path as a continuous stream of bits. This path is maintained for the duration of communication between A and B, and is then released. In packet switching, data is divided into packets (chunks of specific length and characteristics) which are sent from A to B via intermediate nodes. Each intermediate node temporarily stores the packet and waits for the receiving node to become available to receive it. Because data is sent in packets, it is not necessary to reserve a path across the network for the duration of communication between A and B. Different packets can be routed differently in order to spread the load between the nodes and improve performance. However, this requires packets to carry additional addressing information.