Packet Switching
Packet switching is the process by which data is broken into smaller chunks, or packets, for transmission over a network. Packet switching divides a message into smaller units, sends each unit individually via the best route possible at the time, and reassembles the units into the complete message at the destination. Most wide-area network (WAN) protocols are based on packet-switching technologies.
Until the late 1960s computers could only talk to one other computer at a time, using dedicated, circuit-switched communication lines. This was inefficient, as it resulted in brief, intermittent data communications. In 1968 packet switching was first implemented in a government laboratory in England. The idea behind packet switching was that it was a more efficient use of communication time.
Upon hearing about the use of this technology in England, the Department of Defense issued a Request for Proposal (RFP) to implement packet switching for the ARPAnet. Faculty members and graduate students from UCLA responded with a proposal to determine the feasibility and reliability of using packet switching on the ARPAnet. This team designed communication protocols and interfaces to send and receive messages from other computers. By 1969 four university computers were connected via a WAN and exchanging packet switched messages.
The implementation of packet switching requires communications protocols, a router, and network software. TCP/IP, X.25, and frame relay are WAN communications protocols that use packet switching.
Every host connected to a WAN needs a router. A router is a specialized piece of hardware that connects a local-area network (LAN) to a WAN. A router sends packets to other destinations, monitors packets on the WAN for any that are addressed to computers on its LAN, and captures packets addressed to computers on its LAN.
Network software breaks up outgoing messages into packets. Each packet is about one kilobyte in length. The beginning of each packet is called a packet header. The packet header contains the IP address of the source computer sending the packet, the IP address of the destination computer, the length of the packet (measured in bytes), the total number of packets in the complete message, and the sequence number of this packet compared to the other packets comprising the message. Packet header information allows individual packets to be correlated with other packets when they reach the destination computer.
The order of transmission of individual packets is not important, nor is the route that each packet takes to get to its destination. As long as one packet of a message reaches its destination, the receiving computer can determine from the packet header how many packets there were supposed to be and request retransmission of the packets that are missing. When a connection is broken or a computer is down, packets are rerouted across other connections. Also, routers dynamically adjust their internal routing tables according to network conditions to bypass downed or unreliable connections.
This is the complete article, containing 471 words
(approx. 2 pages at 300 words per page).
