Definition[edit | edit source]
Packet switching (also spelled packet-switching) is
|“||[a] system whereby messages are broken down into smaller units called packets that are then individually addressed and routed through the network.||”|
History[edit | edit source]
Packet switching was developed independently in the 1960s by Donald Davies at the National Physical Laboratory in the UK and also for the United States Department of Defense-sponsored ARPANET, the precursor of the modern Internet.
How it works[edit | edit source]
When end users communicate in a traditional “circuit switching” system a dedicated channel is established between them that others cannot use. Using packet switching, on the other hand, computers breaks information into pieces of digital data called “packets” and route them between various nodes on a computer network to a recipient computer that then reassembles the packets into their original form.
The transmission of a single document might involve dozens of packets, each of which might travel a different path through the network to the recipient computer. Each packet contains the address of the recipient and is numbered in sequence and sent out over the network individually. When received at the ultimate destination computer, that computer must re-assemble the packets in the correct order before delivering them to the recipient. Packets may in fact arrive out of order, and quite commonly do because the packets have traveled over quite different physical paths.
One key advantage of packet switching is that computer and telecommunications facilities can be efficiently utilized by interleaving short packets of information from various sources to fill all available transmission capacity. Thus, packet switching allows the sharing of resources at reduced cost. There is a price to pay, however, in terms of delays as the packets await their turns to be delivered. As such, packet switching is extremely well suited to some forms of digital communication, and less well suited to others. The way in which data is processed within the network as generic packets means that different technologies (wireless networks, fiber optic cables, and so on) can be used interchangeably.
Another advantage of packet switching is its resiliency, which is due to the packet’s ability to be transmitted over multiple routes, avoiding areas that may be congested or damaged. Conversely, conventional voice services use traditional [[telephone networks, which are based on circuit switching technology. Instead of breaking a message up into packets, circuit-switching uses a dedicated channel to transmit the voice communication. Once all of the channels are occupied, no further connections can be made until a channel becomes available.
Applications[edit | edit source]
The ideal use of packet switching is for electronic mail. E-mail communication is “asynchronous,” that is, it is not a live or “real time” exchange back and forth between two people. An e-mail message that is divided into packets arriving essentially randomly at the destination computer can therefore be reassembled without a delay that is noticeable to the message’s recipient. Typical “delays” are on the order of fractions of a second in any event, perhaps reaching seconds or minutes during periods of heavy Internet traffic. Most people are unaware and unconcerned whether an e-mail they receive was sent five seconds or five minutes ago.
Some digital communications are not “asynchronous,” however, and hence are better suited to live or real time operation. A telephone call over the Internet is such an example. A phone call is “synchronous” — it is live and in “real time.” When an Internet user places a phone call over the Internet, the call is digitized at the sending end (converted into one’s and zero’s). It is then — like an e-mail message or any other Internet communication — broken up into packets.
Like the packets of an e-mail message, the packets of a “phone call message” may well travel over different routes and may arrive out of order. The receiving computer has no trouble re-arranging the packets into the correct order, but this rearrangement takes some amount of time. Again, this delay may only be fractions of a second to several seconds, but the if resulting delays happen in the middle of a spoken sentence, they introduce some “jerkiness” into the telephone conversation with odd and unpredictable delays for both parties.
It is possible to construct a packet-switched network like the Internet so that even phone calls can be transmitted smoothly, without interruption. But “smooth” transmission of real-time information like phone calls is a function of two things: the network’s capacity and the amount of use it gets. Any given capacity (called “bandwidth”) would allow smooth real-time transmission if the number of users and uses could be limited. Any given number of users and uses could be accommodated if the capacity could be expanded appropriately.