|
A unit of data used to send information
across the internet. Most types of network communications use the
Transport Control Protocol (TCP) layer of the tcp/ ip protocol to
split messages into a number of discrete packets before transmitting them.
Each packet contains the address of its destination and a number denoting
its place in the sequence. Some packets also include information about what
sort of data they are carrying, which allows the network to prioritise them;
for example, sending time-critical video data ahead of text.
Packets
have some significant advantages compared with continuous streams of bits.
packets from the same original message can travel by different routes across
the network, each taking the path that is least crowded at the time of
transmission. Once they have all arrived at their destination they are
reassembled into the correct order. networks that use packets can also
transmit parts of several different messages at the same time. The
disadvantage of packet-based networks is that the process of splitting up
and reassembling messages imposes a time delay and a processing overhead.
The degree to which this matters varies according to the type of packet. ATM
networks, which use a small packet size, have more work to do and thus
impose a greater overhead. |
