Modern automation systems have high reliability requirements, and the delivery of messages in the communication network is time critical. In Ethernet networks where accurate time synchronization is needed, IEEE 1588 is often applied. To satisfy the requirements mentioned before, a network redundancy mechanism that offers zero recovery time in case of failures is needed. Some Ethernet redundancy protocols that fulfill these requirementsoften introduce an abstraction layer that hides the information that the application is running over a redundant network and on which network path a packet was received. This abstraction causes problems if Ethernet redundancy is combined with IEEE 1588, because IEEE 1588 needs exact knowledge of the network path where the messages were received to be able to provide precise synchronization. This thesis analyzes these combination problems, and describes three approaches that can be applied to solve them. Two of these three approaches allow IEEE 1588 to access the timing information of both redundant networks paths. This information will be used in this thesis to improve the synchronization accuracy. Different methods are presented that increase the synchronization accuracy by making use of the additional timing information. A demonstration of the potential for accuracy increase by these methods is given, where the methods are applied on real world measurements. Analyzing the results of this measurement, the average synchronization accuracy was clearly improved. The worst case synchronization error per synchronization round could not be reduced, but the probability of its occurrence could be significantly reduced, which increases the average synchronization accuracy.
