Clock Synchronization Modes

Concepts

Telecommunications services on modern communications networks require that the frequency offset or time difference between devices be within an acceptable range. This requirement must be met using network clock synchronization.

Clock synchronization includes frequency synchronization and phase synchronization.

Frequency synchronization

Frequency synchronization, also called clock synchronization, indicates that signals have the same frequency and a fixed phase difference. That is, signals are sent or received at an average rate. All devices on a communications network operate at the same rate.
Phase synchronization

Phase synchronization, also called time synchronization, indicates that both frequencies and phases of signals are consistent. That is, the phase offset between signals is always 0.

Figure 1 illustrates the difference between frequency synchronization and phase synchronization. If Watch A and Watch B show different time but maintain a constant time difference, for example 6 hours, this is called frequency synchronization. If Watch A and Watch B maintain the same time, this is called phase synchronization.

Figure 1 Differences between frequency synchronization and phase synchronization

Synchronization Modes

Two clock synchronization modes are available on a communications network: pseudo synchronization and master-slave synchronization.

Pseudo Synchronization

In pseudo synchronization mode, each switching site has its own highly accurate and stable clock. The sites do not synchronize their clocks with each other. A difference may exist in the clock frequency and phase among the clocks of the switching sites. However, small differences generally do not affect service transmission and can be ignored. Pseudo synchronization is often used on communications networks between countries. Caesium clocks are usually used in countries.

Master-Slave Synchronization

In master-slave synchronization mode, a highly accurate master clock is deployed on a network and traced by other sites. Other network elements on the network use the master clock as the reference clock. Each sub-site traces the clock of the upper-level site. In this way, the clock is synchronized to endpoints on the network. Master-slave synchronization is typically used on communications networks inside a country or region. Master-slave synchronization is classified as direct master-slave synchronization or hierarchical master-slave synchronization.

Direct master-slave synchronization

Figure 2 shows direct master-slave synchronization. All slave clocks synchronize with the master clock. Direct master-slave synchronization applies to simple networks.
Figure 2 Direct master-slave synchronization
Hierarchical master-slave synchronization

Figure 3 shows hierarchical master-slave synchronization. In this mode, the clocks are classified into three stratum: stratum-1 master clock, stratum-2 slave clock, and stratum-3 slave clock. The stratum-2 slave clock synchronizes with the stratum-1 master clock, and the stratum-3 slave clock synchronizes with the stratum-2 slave clock. Hierarchical master-slave synchronization applies to large and complicated networks.
Figure 3 Hierarchical master-slave synchronization

To improve master-slave synchronization reliability, two master clocks can be deployed: one as the active master clock, and the other as the standby master clock. Under normal circumstances, all network elements, including the standby master clock, trace the active master clock. If the active master clock becomes faulty, the standby master clock takes over as the reference clock for the entire network. After the fault is rectified, the original active-standby relationship is resumed.

Fundamentals

Synchronous Ethernet is a clock synchronization technology that carries and recovers clock frequency information through physical-layer code streams. Figure 4 shows the clock synchronization procedure:

In the transmit direction: Device1 traces the high-precision clock (master clock) and injects the clock information into the PHY chip of an Ethernet interface card. The PHY chip carries the clock information in the serial code stream of the Ethernet line to transmit the information to the downstream node.
In the receive direction: The PHY chip of the Ethernet interface card on Device2 extracts clock information from the serial code stream of the Ethernet line, and then sends the clock information to the clock chip. The clock chip selects the most precise clock (or uses the manually or forcibly specified clock) as the reference clock source based on the automatic clock source selection algorithm and sends the clock source to the system phase-locked loop (PLL). After the PLL traces the reference clock source, the PHY chip of the Ethernet interface card transmits the generated high-precision system clock to downstream devices, and other functions (such as the MAC module) of the local device use clock signals.

Figure 4 Synchronous Ethernet
click to enlarge

Comparisons between synchronous Ethernet and other clock synchronization protocols

Clock Protocol	Whether Frequency Synchronization Is Supported	Whether Time Synchronization Is Supported	Time Synchronization Accuracy	Signal Transmission Mode
NTP	No	Yes	Millisecond accuracy	Time signals are transmitted using NTP packets.
Synchronous Ethernet	Yes	No	-	Clock signals are transmitted using serial data streams at the physical layer, without affecting upper-layer services and CPU performance.
PTP	Yes	Yes	Sub-microsecond accuracy	Clock and time signals are transmitted using PTP packets, and higher time accuracy is achieved with the assistance of hardware.