Formal analysis of transient transmission errors in Ethernet networks considering message timing.

What is it about?

The paper starts by providing an approach for the calculation of the probability of a transient error during the transmission of a message in an Ethernet network, hereby assuming a known and constant bit error rate (BER) and the freedom from interference, that is no jitter caused by other messages in the network. After this, we show that, considering the system model adapted in many safety-critical systems, automotive in particular, where a transmission is recognized as a failure if no correct messages are received within a predefined time interval, the timing of messages within the communication network has to be considered during the calculation of the worst-case failure probability. We address this issue by proposing an approach that combines the probability calculations with a state-of-the-art approach for the timing analysis of messages in Ethernet networks to derive a safe upper bound for the probability of transient transmission failures. In the experiment section, we show how adjusting the sending rate of messages can improve the reliability and highlight the correlations between timing and transmission reliability in Ethernet networks. The experimental results provide evidence that timing and reliability of communication networks have to be considered jointly.

Featured Image

Why is it important?

This paper a) shows that neither the timing nor the reliability of message transmissions should be optimized in isolation, b) proposes the adjustment of the sending rates as a novel, cost-efficient mechanism to improve the transmission reliability and c) provides an approach for the reliability analysis that considers the interrelations between message timing and the probability for a transient failure.