Where transport protocols are needed and where not? In the field of on-board communications transport protocols are not used because the information to exchange the control devices while driving, usually short. That is, the signals are smaller than the data length of CAN, FlexRay and LIN messages. So you can use these messages directly at the level of the data link layer. The fact that one uses no transport protocol, to reduce the latency and thus the transmission time. Typically are packed even more of these signals in a message to minimize the overhead on the wire. This is the responsibility of an application layer, as is provided for example by OSEK or Autosar. It is called the interaction level.

In the area of ​​off-board communication, there are also several short blocks of data, for example for the retrieval of data such as the engine speed. Data blocks of medium length that do not fit everything into a single CAN message are, for example, to query the frame numbers. Finally, there are also, for example when flashing long data blocks that are to be accommodated, even a long FlexRay message no longer in a single message. Now it is up to the transport layer to adjust the message length. At the application level, the messages can be as long as is specified on the bus level, the maximum message length of the bus system. The task of the transport protocol is to split long messages into short and the inverse process on the receiving end, the assembly of short bus messages to a long diagnostic message. This is known as segmentation and Desegmentierung.

If a large transfer large data blocks segmented can lead to the use of simple control devices to ensure that they get too many messages they can not be processed immediately. There is a so-called flow control or flow control, where transmitters and receivers to communicate as to when the messages are sent.