TCP/IP model

The DoD Internet Architecture Model is a layered architecture for network protocols.

In this model, the individual tasks are divided in data transmission in successive layers (see also TCP / IP reference model). There are a number of protocols for each layer (see Internet protocol suite ), which solve the tasks of each layer in different ways. The model was in the late 1960s by DARPA for the Defense of the United States ( Department of Defense, in short: DoD) developed and initially consisted of four layers. It was about a purely military application in the context of the development of the ARPANET, which should be protected by a decentralized structure against failures.

Layer

Network Access layer

In the network access layer, the problem of data transmission between computers and processes is achieved at the physical level. This is of importance, which transmission medium ( copper cable, wireless network, etc. ), which line code ( RZ code, NRZ code, AMI - code, etc. ) and which access protocol (eg, Ethernet, V.24) will be used. With the help of an address is specified here, which machine or process is to receive the data packet.

Internet layer

The task of the protocols of the Internet layer is to provide a network-wide addressing independently eg the transmission medium allow. Any computer or any network card is assigned a network address. Unlike the address of the Network Access layer, this address is assigned according to logical criteria and thus enables large networks can be managed. It is similar to phone numbers to which a prefix belongs and thus prevents a switching system must know all the phone numbers in the world. Thus, it is possible to send packets over multiple networks. Most Widespread protocol of this layer is the Internet Protocol.

Host - to-Host layer

The task of the protocols of the host-to- host layer is to allow the transport of data between two processes on different computers is ( end-to -end connection ). This is necessary because it is not sufficient to transport the data to the target computer, but it must also be stated that the ongoing programs there that are waiting for data from the network, to receive each package. There are different approaches to information transfer:

• In a connectionless transmission in this layer no control is installed so that the package will be delivered. The inspection must be carried in the Process layer. Similar is with the sending of a letter in the real world.
• With a confirmed, connectionless transmission, a confirmation message is provided to report the receipt of the package the transmitter. It corresponds to the sending of a registered letter with acknowledgment of receipt.
• In a connection-oriented transmission before the transmission is a logical connection is established and maintained. Even when no data is transmitted, control information is exchanged between the computers. This method allows control of the packet sequence and data flow control.

On the Internet, this is what the two TCP and UDP protocols available. When TCP is a connection-oriented protocol. It allows for a reliable data transmission. However, this also means a lot of extra information that is transmitted in addition to the actual message. The UDP on the other hand is a connectionless protocol. Due to the reduced administrative burden the network traffic for this protocol is kept low, and there may be a higher data throughput can be achieved. For UDP does not recognize independently the loss of packets, and there is a danger that all messages arrive in a different order than they were sent.

Process layer

The minutes of the Process layer define the structure of the actual message. Here is the actual user data are stored according to the protocol chosen. The best known are used to transfer e-mail (SMTP), web pages (HTTP) or file transfer (FTP ). For any information that needs to be transferred, where their exact position in the embassy and their precise structure is fixed.

Packet structure

Each level of the layer model adds while sending the data packet header information as which will be evaluated by the receiver at the same level. One can think of it as a large envelope in which a slightly smaller envelope attached and inside this a very small envelope, which ultimately contains only the leaf with the actual information. Each envelope while the information is required in each layer:

The OSI reference model

Some years after the DoD model, the OSI model by the International Organization for Standardization ( ISO). The structure is similar to the DoD model, however, instead of just over seven to four layers. Although the DoD model had proven its effectiveness in practice, considering the large computer companies to use the OSI model for their own communication protocols. Since 1988, even the U.S. government supported the new model. Nevertheless, the four-layer principle did eventually through. However, the OSI model is often used to boot to describe communication processes on the Internet. For this it is necessary to summarize some layers of the OSI model: