This section describes network synchronization capabilities available on SR OS platforms. These capabilities involve multiple approaches to network timing; namely SDH/SONET, Synchronous Ethernet, BITS, and Adaptive clocking and a Precision Time Protocol (PTP) IEEE 1588v2. These features address barriers to entry by:
Providing synchronization quality required by the mobile space; such as radio operations and circuit emulation services (CES) transport.
Augmenting and potentially replacing the existing (SONET/SDH) timing infrastructure and delivering high quality network timing for time sensitive applications in the wireline space.
Network synchronization is commonly distributed in a hierarchical master-slave topology at the physical layer as shown in Figure 1.
The architecture shown in Figure 1 provides the following benefits:
Limits the need for high quality clocks at each network element and only requires that they reliably replicate input to remain traceable to its reference.
Uses reliable physical media to provide transport of the timing signal; it does not consume any bandwidth and requires limited additional processing.
The synchronization network is designed so a clock always receives timing from a clock of equal or higher stratum or quality level. This ensures that if an upstream clock has a fault condition (for example, loses its reference and enters a holdover or free-run state) and begins to drift in frequency, the downstream clock is able to follow it. For greater reliability and robustness, most offices and nodes have at least two synchronization references that can be selected in priority order (such as primary and secondary).
Further levels of resiliency can be provided by designing a capability in the node clock that operates within prescribed network performance specifications without any reference for a specified time-frame. A clock operating in this mode holds the last known state over (or holdover) until the reference lock is again achieved. Each level in the timing hierarchy is associated with minimum levels of network performance.
Each synchronization capable port can be independently configured to transmit data using the node reference timing or loop timing. In addition, some TDM channels can use adaptive timing.
Transmission of a reference clock through a chain of Ethernet equipment requires that all equipment supports Synchronous Ethernet. A single piece of equipment that is not capable of performing Synchronous Ethernet breaks the chain. Ethernet frames still get through but downstream devices should not use the recovered line timing as it is be traceable to an acceptable stratum source.