A model-based time distribution implementation for asynchronous systems

Precise time distribution amongst electronic systems within the engineering industry is a challenging yet necessary objective to ensure successful operations. Critical decisions are controlled by computers based upon sensor and other instrumentation data inputs. The collected data's source will append a timestamp to each message which is used by the computer in order to determine whether the data is timely or stale. These timestamps are also used to determine when major and minor events occur including video playback. The system level time distribution implementation must be decided in the early phases of a program life cycle in order to provide an affordable solution to the customer and the decomposed requirements flowed down to the vendors are achievable to meet the design goal. Validation of the chosen method proves difficult when the electronic system is not physically integrated. This paper describes in detail a simulation tool that was created to model the internal time accuracy of a large scale electronic system's time accuracy relative to an external time source. The model accounts for multiple variables of internal electronics time errors in asynchronous systems such as propagation delay, time server accuracy, internal processing delays, transmission delays, clock stability, and the number of electrical components connected in the time distribution layers. The user can input these accuracy parameters into the model to develop worst case time accuracy scenarios. There are two simulation scenarios discussed in the paper, a system utilizing an NTP server and a GPS server. One can then visualize, during operation, the behavior of each electrical system's asynchronous time distribution implementation.