Precision timing plays a critical role in the synchronization of telecommunications networks. It enables service providers to use available spectrum resources more efficiently, increase bandwidth and deliver high-speed wireless services. Until now, service providers have used timing sources based on the Global Navigation Satellite System (GNSS) or Global Positioning System (GPS) to provide frequency and time/phase synchronization; but this is changing. Some of the drivers for change relate to impediments associated with GNSS/GPS itself. As well, the sheer growth of wireless network sites required for 5G will make alternate sources of primary or secondary backup timing necessary to keep site costs down and ensure high availability.
To meet demand for high capacity, service providers need to densify their networks with new 5G radios. 5G can be deployed in a wide range of settings, from outdoors to indoors, in tunnels, stadiums, buildings and subways. As the number of sites increases, the costs to have GNSS/GPS receivers everywhere also increase.
To reduce infrastructure costs, some RAN equipment manufacturers are removing GNSS/GPS receivers from the radio, relying instead on network-based timing to obtain the synchronization source. In doing so, they are also mitigating risks associated with GNSS/GPS receivers, which are vulnerable to spoofing, jamming and line-of-sight limitations. They also eliminate issues such as interference caused by weather and other factors.
Many new 5G radio deployments will use Time Division Duplex (TDD) spectrum bands with larger swaths of spectrum to increase throughput. In addition to frequency synchronization, these 5G TDD implementations need time and phase synchronization to operate. This imposes stricter reliability requirements on the network due to holdover budget. GNSS/GPS failures may impact RAN performance starting within hours from the time of initial failure, versus days, in the case of sites needing only frequency synchronization.
This creates the need for higher-grade and more expensive oscillators, such as double oven-controlled crystal oscillators (DOCXOs) or rubidium oscillators. While too costly to have these oscillators at each cell site, they can be placed at hub locations where network-based timing is used to distribute holdover synchronization. To keep costs in check, these oscillators should serve as many cells as possible to minimize the number of oscillators required. Furthermore, by using network-based timing, operators can benefit from better network redundancy and protection. Operators can use the same equipment they install for redundancy to support synchronization, thereby lowering the cost and complexity of cell sites.
5G also introduces more flexibility into the architecture to meet more diverse service requirements. 5G allows flexible placement of RAN functions such as the centralized unit (CU) and distributed unit (DU). They can now be separated from the radio units (RUs) at the cell site and located more centrally as in Cloud RAN implementations. This disaggregation of the RAN makes synchronization even more challenging because the stringent absolute and relative timing requirements (which can demand accuracy in the tens of nanoseconds) must be met. This creates the need for tight connectivity and highly precise synchronization distribution within fronthaul transport networks because the components must operate as if they are collocated to maintain processing integrity.
To improve bandwidth efficiency, 5G radios will use the new eCPRI protocol, which reduces bandwidth requirements tenfold compared to CPRI. Unlike CPRI, however, the eCPRI protocol is not designed intrinsically to carry synchronization information. This means packet-based synchronization protocols such as PTP are required to provide synchronization.
The ITU-T has defined a new standard (G.8273.2) that adds enhanced boundary clocks (Class C and D) to meet the stringent synchronization requirements of disaggregated 5G networks, particularly as they apply in fronthaul networks. Telecom boundary clocks (T-BCs) allow accurate distribution of timing in the network, for example, by using Time Sensitive Network (TSN) Ethernet bridges having T-BC functionality.
A boundary clock recovers time information from the PTP messages that it receives from an upstream node and uses this information to update its internal time counters. It then generates new PTP messages to provide timing information to a downstream node. By accounting for the residence time of packets within a node, the boundary clock recovers time information with the highest accuracy. T-BCs incorporated into transport equipment enable timing distribution across the synchronization trail.
The ITU-T has also added a new enhanced Ethernet equipment clock (eEEC) standard (G.8262.1). It defines performance requirements for new synchronous equipment end clocks. Improved frequency synchronization is vital because it helps to syntonize the network and leads to more accurate phase/time synchronization. When combined with PTP, eEEC enables accurate time recovery through PTP that meets Class C and Class D specifications for time recovery.
Network-based timing also enables improved visibility of flows. With network-based synchronization using PTP, it is possible to monitor the flow from the grandmaster clock to the PTP client. This gives the service provider a complete view of all synchronized PTP clients, which increases visibility and control of the network.
As 5G networks continue to be rolled out, we will see operators increasingly turn to network-based timing as either a primary or secondary backup source of synchronization. The reasons for doing so are clear. It offers operators the ability to reduce cell site costs, mitigate the impediments to GNSS/GPS-based timing, improve protection and availability, and provide better visibility of synchronization flows, thus letting operators reap the full benefits of 5G.
