Error Detection Capabilities of Automotive Network Technologies and Ethernet - A Comparative Study

Coding the payload data and sending the codeword as an overhead of the packet is a very common way to protect data in communication networks. The protection level of the coding technique is chosen depending on the importance of the transmitted data. While high priority, safety critical applications do not tolerate any single bit error in data packets, lower priority services can handle few bit errors. Coding techniques provide error detection and up to a certain level also error correction. Bit errors that are not detectable by coding techniques at the receiver side occur with a residual error probability. The better the coding technique is, the lower is its residual error probability for different bit error rates. Most automotive network systems use cyclic redundancy codes (CRC) mainly in order to detect transmission bit errors. Instead of correcting the identified bit errors which is quite time consuming, usually a retransmission of the damaged data packet is triggered. Similar to automotive network systems, Ethernet, the most applied network technology in local area networks uses the CRC error detection technique. In this work, we present a comparative study of the error detection capabilities of automotive network systems and Ethernet as a possible network system for time critical applications in the car. We evaluate the related residual error probabilities for a reasonable range of bit error rates. Furthermore, several commercial concepts are presented from the automation field that increase the error detection capability of the standard Ethernet technology significantly.