packet switching

packet switching

Packet switching

A software-controlled means of directing digitally encoded information in a communication network from a source to a destination, in which information messages may be divided into smaller entities called packets. Switching and transmission are the two basic functions that effect communication on demand from one point to another in a communication network, an interconnection of nodes by transmission facilities. Each node functions as a switch in addition to having potentially other nodal functions such as storage or processing.

Switched (or demand) communication can be classified under two main categories: circuit-switched communication and store-and-forward communication. Store-and-forward communication, in turn, has two principal categories: message-switched communication (message switching) and packet-switched communication (packet switching).

In circuit switching, an end-to-end path of a fixed bandwidth (or speed) is set up for the entire duration of a communication or call. The bandwidth in circuit switching may remain unused if no information is being transmitted during a call. In store-and-forward switching, the message, either as a whole or in parts, transits through the nodes of the network one node at a time. The entire message, or a part of it, is stored at each node and then forwarded to the next.

In message switching, the switched message retains its integrity as a whole message at each node during its passage through the network. For very long messages, this requires large buffers (or storage capacity) at each node. Also, the constraint of receiving the very last bit of the entire message before forwarding its first bit to the next node may result in unacceptable delays. Packet switching breaks a large message into fixed-size, small packets and then switches these packets through the network as if they were individual messages. This approach reduces the need for large nodal buffers and “pipelines” the resources of the network so that a number of nodes can be active at the same time in switching a long message, reducing significantly the transit delay. One important characteristic of packet switching is that network resources are consumed only when data are actually sent.

All public packet networks require that terminals and computers connecting to the network use a standard access protocol. Interconnection of one public packet network to others is carried out by using another standardized protocol.

Packet-switched networks using satellite or terrestrial radio as the transmission medium are known as packet satellite or packet radio networks, respectively. Such networks are especially suited for covering large areas for mobile stations, or for applications that benefit from the availability of information at several locations simultaneously.

Asynchronous transfer mode (ATM) is a type of packet switching that uses short, fixed-size packets (called cells) to transfer information. The ATM cell is 53 bytes long, containing a 5-byte header for the address of the destination, followed by a fixed 48-byte information field. The rather short packet size of ATM, compared to conventional packet switching, represents a compromise between the needs of data communication and those of voice and video communication, where small delays and low jitter are critical for most applications.

Data communication (or computer communication) has been the primary application for packet networks. Computer communication traffic characteristics are fundamentally different from those of voice traffic. Data traffic is usually bursty, lasting from several milliseconds to several minutes or hours. The holding time for data traffic is also widely different from one application to another. These characteristics of data communication make packet switching an ideal choice for most applications. The principal motivation for ATM is to devise a unified transport mechanism for voice, still image, video, and data communication. See Data communications

packet switching