Telecommunications services provided by telcos and long distance carriers that route packets of data between local area networks (LANs) in diverse geographical locations to form a wide area network (WAN). Packet-switching services are used to connect multiple LANs into a point-to-multipoint configuration, usually called a multipoint WAN.
A customer’s local network is typically connected through routers, bridges, frame relay access devices (FRADs), or other devices to a telco’s central office (CO). These devices either have built-in technology for connecting directly to packet-switching services or use intermediary devices located at the customer premises. Packet-switching devices take network frames and “package” them into packets suitable for the type of packet-switching service being used. These services can include frame relay, X.25, Asynchronous Transfer Mode (ATM), or Switched Multimegabit Data Services (SMDS) public or private packet-switched networks. If virtual private network (VPN) technologies are used, the public Internet can also be used as a packet-switching service for multipoint WAN connections.
Graphic P-3. Packet-switching services.
The packaging process varies with the particular service used, but it basically consists of breaking down network frames into relatively small individual packets of data and tagging the packets with the destination address of the remote node to which the packet is directed. Each end node (local network access device) connected to the cloud has a layer 2, or data-link layer, address that is known to every other end node. These addresses are used to route packet data between individual nodes on the WAN or to broadcast packets to all nodes when needed. Other information is also tagged onto the packets for error correction and other purposes, depending on the service used. The packets are usually small to lessen the load on the switching devices and to enable quick retransmission when transmission errors occur.
Packets are individually placed onto the carrier’s packet-switched network and switched from circuit to circuit until they reach their destination. Two packets forming part of the same network message might take entirely different routes to reach their destination node - it depends on the best route available at any given moment, as determined by the packet-switching services themselves. This is different from circuit-switched networks, in which all packets are sent over the same switched circuits for the duration of the connection. At the destination, the packets are reassembled into network frames and delivered to the remote network, where they are routed to their destination computers.
In networking diagrams, a public packet-switched network is typically depicted as a cloud because the details of the switches and connections are not of interest to the customer - they are the responsibility of the carrier or carriers providing the services.
Advantages of packet-switching services include the following: