Internet Protocol

from Wikipedia, the free encyclopedia
IP (Internet Protocol)
Family: Internet protocol family
Operation area: Sending data packets
both locally and
worldwide via various
networks
IP in the TCP / IP protocol stack :
application HTTP IMAP SMTP DNS ...
transport TCP UDP
Internet IP ( IPv4 , IPv6 )
Network access Ethernet Token
bus 		Token
ring 		FDDI ...
Standards: RFC 8200 (IPv6, 2017)
RFC 791 (IPv4, 1981)

The Internet Protocol ( IP ) is a network protocol that is widespread in computer networks and forms the basis of the Internet . It is the implementation of the Internet layer of the TCP / IP model or the network layer of the OSI model . IP is a connectionless protocol, i.e. no status is established for the communication partners.

Features and functions

The IP forms the first layer of the Internet protocol family that is independent of the transmission medium . This means that by means of IP address and sub net mask (subnet mask) for IPv4 , or prefix in IPv6 , computer within a network into logical units called subnets , can be grouped. On this basis it is possible to address computers in larger networks and to send them IP packets , since logical addressing is the basis for routing (route selection and forwarding of network packets).

Address allocation

As a rule, public IP addresses must be uniquely assigned worldwide, which is why their assignment is regulated by the Internet Assigned Numbers Authority (IANA). This delegates large address blocks to the Regional Internet Registries (RIRs), which then assign subnets thereof to Local Internet Registries (LIRs). The LIRs include, for example, Internet providers who assign smaller subnets or individual addresses to customers from their address range.

With IPv4, the address space to be allocated is largely used up. In February 2011 the IANA assigned the last address blocks to the RIRs.

history

In May 1974, Vint Cerf and Bob Kahn published a research paper on a network protocol for cross-communication between different packet-switched networks. In the model lead terminals ( English hosts ) a "Transmission Control Program" ( English transmission control program - TCP) from which the transmission of a continuous data stream between processes ensures. Gateways take over the transformation of packets at network boundaries.

The first complete protocol specification was released with RFC 675 in December 1974. The monolithic Transmission Control Program was later divided into a modular architecture that made the Internet Protocol ( english internet protocol ) for host-to-host communication and the transmission control protocol ( English transmission control protocol - TCP) for process-to-process communication. The model became known as the TCP / IP reference model .

Both protocols were revised several times before they were put into practical use. In addition to the final designation as " Internet Protocol ", " Internetwork Protocol ", " Internet Datagram Protocol " or " DoD Standard Internet Protocol " were also used in drafts . In the event of major changes to the IP header , a version number contained in the header was incremented. When TCP / IP was introduced in the ARPANET on January 1, 1983, IP packets were therefore version number 4. Previous versions were not widespread.

In the first draft of the protocol, an addressing scheme of variable length was provided, consisting of a network address at least 4 bits long, a 16-bit host address and a 24-bit port number. Later, IP addresses were set to 32 bits, consisting of an 8-bit network address and a 24-bit host address. The port number was moved to TCP and shortened to 16 bits. With RFC 791 , network classes were introduced in order to have more flexibility in dividing an IP address into network and host parts. Subnetting was not planned at the time. Jon Postel took care of the assignment of network addresses - a role that was later known as the Internet Assigned Numbers Authority .

With the emerging scarcity of IP addresses, the development of a successor protocol began in the early 1990s. To differentiate, the established Internet protocol was designated as IPv4 and the new Internet protocol as IPv6 according to the version number in the IP header . The most important innovation is the considerably larger address space: compared to the 32-bit addresses in IPv4 (results in approx. 4 billion, or 4.3 · 10 9 addresses), IPv6 uses 128-bit addresses (results in approx. 340 sextillion, or 3 , 4 x 10 38 addresses).

Version number 5 was occupied by the experimental Internet Stream Protocol , which was not intended as a successor, but rather as a supplement to the Internet Protocol. The Internet Stream Protocol was later abandoned without having achieved any notable widespread use. The version numbers 7 to 9 were used for various proposals for an IPv4 successor, but these were abandoned in favor of IPv6.

The spread of IPv6 is slowly increasing, but it is lagging behind the spread of IPv4. Common operating systems and standard software support both protocols. Transition mechanisms enable the simultaneous operation of IPv4 and IPv6 on the same infrastructure. Well-known websites and Internet providers have been offering IPv6 since World IPv6 Day and World IPv6 Launch Day in 2011 and 2012 .

reliability

The design principles of the Internet protocols assume that the network infrastructure at each individual network element or transmission medium is inherently unreliable. These also assume that the infrastructure behaves dynamically with regard to the availability of connections and nodes. However, in order to maintain the network infrastructure, the main focus of data transmission is intentionally largely placed on the end node of each individual data transmission . Routers in the transmission path only send data packets to directly accessible and known transitions that compare the addresses specified for the destination from the route planning prefix.

As a result, these Internet protocols only provide the best possible transitions, which means that these services are characterized as unreliable. The IP is connectionless, each individual data packet is handled independently. Since each individual transmission path of a data packet is redefined (dynamically), it is possible that the packets are sent to their destination on different paths.

Internet protocol version 4 (IPv4) provides the necessary protection to ensure that the protocol header of every data packet is error-free. A route planning node calculates a checksum for the packet header. If the checksum is invalid, the route planning node discards the packet. The route planning node does not need to advertise an end node, although the Internet Control Message Protocol (ICMP) allows such advertisements. In contrast, Internet Protocol version 6 (IPv6) does not have a checksum, which leads to faster processing during route planning.

All sources of error in the transmission network must be discovered and replaced with the help of the transmission to end nodes. The upper layer protocols of the Internet protocol family are responsible for resolving reliability problems. For example, a host can withhold data and perform a correction before the data is delivered to the intended recipient.

Link capacity and performance

Even if the transmission path is available and reliable, because of the dynamic nature and heterogeneity of the Internet and its components, there is no guarantee that each of these individual paths will in fact be able to carry out a data transmission. For example, the permitted transmission size of the respective data packets represents a technical restriction. Every application must ensure that the correct transmission properties are used.

Part of this responsibility also lies in the upper-tier protocols. IPv6 uses the ability to examine the maximum transmission unit size of a local connection as well as the completely planned path to the destination. The IPv4 interconnection layer originally had the ability to automatically break down large data packets into smaller units for transmission.

The Transmission Control Protocol (TCP) is an example of a protocol that regulates its segment size to be less than the maximum allowable flow, the Maximum Transmission Unit (MTU). However, the User Datagram Protocol (UDP) and Internet Control Message Protocol (ICMP) ignore the MTU size, forcing the IP to split oversized data packets.

See also

related protocols

Web links

Individual evidence

  1. ^ Vint Cerf, Bob Kahn: A Protocol for Packet Network Intercommunication . In: IEEE Transactions on Communications . tape 22 , no. 5 . IEEE, May 1974, ISSN  0090-6778 , doi : 10.1109 / TCOM.1974.1092259 (English, ieee.org ).
  2. ^ A b Vinton Cerf, Yogen Dalal, Carl Sunshine:  RFC 675 . - Specification of Internet Transmission Control Program . December 1974. (English).
  3. ↑ Jonathan B. Postel: Draft Internetwork Protocol Specification - Version 2. IEN February 28, 1978, accessed February 9, 2020 .
  4. ↑ Jonathan B. Postel: Internetwork Protocol Specification - Version 4. IEN June 41, 1978, accessed February 9, 2020 .
  5. ↑ Internet Datagram Protocol - Version 4. IEN 80. Jon Postel, February 1979, accessed February 9, 2020 .
  6. ^ DoD Standard Internet Protocol. IEN 123. Jon Postel, December 1979, accessed February 9, 2020 .
  7. ↑ a b Jon Postel (Ed.):  RFC 760 . - DoD Standard Internet Protocol . January 1980. (English).
  8. ^ J. Postel:  RFC 801 . - NCP / TCP transition plan . November 1981. (English).
  9. ^ Jon Postel (Ed.):  RFC 791 . - Internet Protocol . September 1981. (English).
  10. ↑ RFC 1190 . (English).
  11. ↑ Version Numbers. IANA, November 6, 2018, accessed February 9, 2020 .
  12. ↑ a b packet-based communication protocols, Harald Orlamünder, Hüthig