5G Time Synchronization – Does anybody really know what time it is?

The economics of building out a 5G infrastructure is quite different than previous generations of cellular networks. This shouldn’t be surprising given that the goals of 5G include a 10× improvement in user experienced data rate, latency, and connection density. One of the primary architectural changes is a large dependency on small cells. But expensive hardware that makes sense at a small number of macro cell sites doesn’t make sense when a large number of small cells is deployed. One important example is a highly accurate (and expensive) GPS receiver for synchronization. Small cells also have the problem that many of them might be placed in locations that can’t actually see GPS satellites.

All cellular architectures require frequency synchronization. This can, in many cases be achieved through the use of Synchronous Ethernet in the Backhaul and Fronthaul networks. But this is not always feasible, and it doesn’t solve the problem of phase and time synchronization. Several cellular technologies require phase or time synchronization, and time synchronization is also needed for geolocation. The solution in these cases is the Precision Time Protocol (also known as IEEE 1588).

PTP has protocol profiles (parameter sets) for providing frequency synchronization and time synchronization. Since PTP is a networking protocol the accuracy of the synchronization is heavily dependent on a well-designed and controlled network. Network congestion, software and interrupt latencies in PTP servers and client, and similar impairments will affect the accuracy of the synchronization.

PTP has two primary types of nodes (called clocks). Master clocks have an accurate clock source such as a GPS, or atomic clock, or a higher-level master clock. Slave clocks are the nodes that need the synchronization. For scalability and resiliency there can be multiple master clocks that can synchronize (i.e., act as a slave clock) to higher level grandmaster clocks and then provide synchronization to other slave clocks. These are called boundary clocks. So, grandmaster, master, and slave clocks form a hierarchy for synchronization. One other type of clock is a transparent clock. Transparent clocks can be implement in devices (often switches or routers) between master and slave clocks and can make small adjustments in the timestamps of PTP messages that flow through them to account for delay variation in the switch or router, thereby improving the accuracy of the synchronization.

PTP will be an important technical component in 5G networks and an important economic solution for 5G deployment. But synchronization is a tricky business, and without a carefully thought-out design and implementation it won’t provide a satisfactory solution in either realm.