Packet switching

from Wikipedia, the free encyclopedia

The packet switch ( English switching packet ) is a method for data transmission in computer networks . Longer messages are divided into individual data packets and either transmitted without a connection (as a datagram ) or via a virtual connection .

A package typically contains

  • the source of the package,
  • the destination of the package,
  • the length of the data part,
  • the package serial number
  • the classification of the package and
  • the data part

In packet switching, the packets traverse the network as independent and autonomous units and can be temporarily stored in the switching nodes. This is a major advantage, since the transmission speed between individual sections is no longer a limit. However, the architecture appears as a network of queues . Each network node to be passed receives the packet and forwards it to its output interface (the network node), which can, however, be the destination of many transmissions, so that there is a tendency towards overload situations. The existence of the queues creates delays in the delivery of packages or package losses, which can result in a new transmission and thus a further load on the transmission path. This process is not transparent for the user. He does not need or receive any information about the transmission path. As a rule, he cannot obtain this information either, since the transmission paths change dynamically.

The system was developed by the English IT pioneer Donald Watts Davies and Leonard Kleinrock in the USA.

Examples of packet-switched networks are the Internet or modern cellular networks that use standards such as UMTS .

Packet switching methods

A distinction is made between two packet switching methods:

  • Connectionless packet switching (based on datagram messages)

In connectionless packet switching, each packet is provided with a destination address, source address and port numbers ; it can also be identified with a relevant sequence number. This eliminates the need for an exact path (directions to the destination recipient). It does mean, however, that much more information than necessary must be stored in the packet header. Every package is sent and can get to its destination in different ways. As a result, every system now has to do a lot more work than in return for connection-oriented transmission. At the destination, the original message / data is reassembled in the correct order, based on the packet sequence number. This virtual connection, also known as a virtual circuit or byte stream, is provided to the end user through a transport layer protocol.

Connectionless protocols are e.g. B. Ethernet , IP and UDP (User Datagram Protocol) .

  • Connection-oriented packet switching (virtual circuit switching)

In contrast to connectionless packet switching, in connection-oriented packet switching, each packet is provided with a connection ID instead of an address. This address information is only transmitted to each individual node during the connection setup phase. If this information is addressed on the way to the destination receiver, an entry is made in the switching tables of each relevant network node and the connection is switched. The transmission or forwarding of such a packet is very simple, since it is only necessary to "look up" in the switching tables of the nodes in order to determine the relevant connection ID. The packet header required for this contains much less information (ID, length, time stamp or sequence number).

Connection-oriented protocols are e.g. B. TCP, X.25, Multiprotocol Label Switching (MPLS) and Asynchronous Transfer Mode (ATM).

advantages

Compared to circuit switching, packet switching offers a number of advantages.

  • More efficient utilization, since a line is not used exclusively, but several users or services can communicate at the same time.
  • The resources can be distributed fairly among the participants (see also net neutrality ).
  • If there are several routes from the sender to the receiver, the data stream can be transparently rerouted if a switching station fails.

disadvantage

  • Since the transmission paths are not defined, individual switching stations can be overloaded.
  • The packets can arrive at the recipient in a different order than they were sent (e.g. through different transmission paths).
  • A constant data rate is not guaranteed, i. H. the data rate can fluctuate.

Packet switched networks

Individual evidence

  1. ↑ packet switching. In: Dimaweb-Interlexikon. (Technical terms), dimaweb.at, January 25, 2016.
  2. ↑ M. Smouts: Packet Switching evolution from narrowband to broadband ISDN. University of Michigan, Artech House 1992, ISBN 0-89006-542-X .

Web links