Reproducible IEEE 1588-performance tests with emulated environmental influences

The IEEE 1588 standard is widely established and accepted for clock synchronization in Ethernet networks. High accuracy IEEE 1588 implementations require a well-coordinated hardware/software co-design. Processing of sophisticated control algorithms for the time speed of local clocks within synchronization slaves are particularly important. Hence, IEEE 1588 implementations should be tested with respect to certain performance indicators like control loop behavior and synchronization accuracy. Furthermore, interoperability with other devices and standard compliance also need to be considered. The achievable synchronization accuracy depends on environmental conditions, network load and network topology. A test system should be able to emulate these physical conditions. This requires a well-founded knowledge about influences on IEEE 1588 implementations caused by both, environmental conditions and network load. This could be a frequency drift of crystal caused by either temperature variations or mechanical stress. Unfortunately, the standard does not specify an expected behavior in such an environment. Due to lack of standardization and test methods, a system-wide guarantee for synchronization accuracy can only be given for proprietary closed systems. In this paper a reproducible test environment with the ability to emulate environmental conditions is presented, followed by an evaluation of two exemplary implementations. The possibility to guarantee synchronization accuracy with the help of appropriate certification tests in such a specific test environment will be demonstrated.