Packet switching is a digital network communication process in which data is divided into packets that are independently routed from source to destination through shared network links. Packet switching is a key technology in modern telecommunications networks as it allows for asynchronous communication in which data is divided into smaller segments, each with its own addressing information, and then delivered via shared networks rather than dedicated circuits.
Packet switching differs from circuit-switching where dedicated communication paths between devices are established before the transfer of data. Packet switching is used in wide-area networks (WANs), local-area networks (LANs) and metropolitan-area networks (MANs), and may also be included in home-networking technologies. A packet-switched network is composed of two different logical divisions: the physical layer, which is responsible for signal and data transmission across the physical medium, and the network layer, which is responsible for routing and packet forwarding.
Packet switching was first developed in the late 1960s by Paul Baran and is closely related to datagram and virtual circuit technologies. It is the primary technology employed in the Transmission Control Protocol/Internet Protocol model of networking, which is the basis of modern communications networks. Packet switching helps increase network bandwidth and reduces transfer delays by allowing network traffic to be shared across multiple paths, rather than relying on a single route.
Packet switching is also advantageous in networks with widely varying traffic patterns as packet transmission is more efficient than a circuit-switched architecture. This is because idle resources are not dedicated to any particular connection. Additionally, packet-switched networks are more resilient, as they can quickly and easily switch traffic to alternate paths if a failure is detected.
