Cisco 3845 Software Guide - Page 130

Understanding SyncE Chapter 6 Configuring Synchronous Ethernet The reference clock source can be: • A Building Integrated Timing Supply (BITS) clock input • A PHY-recovered clock from uplink ports. The ME 3800X and 3600X switch supports a PHY-recovered clock only from the small form-factor pluggable (SFP+) uplink ports with 10 Gigabit SFP+ or 1000BASE-X fiber SFP modules. All uplink and downlink ports transmit data on the same reference clock. The switch monitors each input clock for frequency accuracy and activity. An input clock with a frequency out-of-band alarm or an activity alarm is invalid. Invalid clocks are not selected as the reference clock. During normal operation, the reference clock is selected based on an algorithm that uses the priority rankings that you assign to the input clocks by using the network-clock-select priority priority global configuration command. Priority 1 is the highest, and priority 15 is the lowest. If you try to assign the same priority to more than one clock, error message appears. Unused input clocks are given a priority value of 0, which disables the clocks and makes them unavailable for selection. The clock selection is based on signal failure, priority, and manual configuration. If you have not manually configured a reference clock, the algorithm selects the clock with the highest priority that does not experience signal failure. With this configuration, pure priority-based mode, an intermittent failure or changes in the network topology can cause timing loops or a loss of connectivity with the clock reference. The Ethernet Synchronous Messaging Channel (ESMC) with source-specific multicast (SSM) provides a way to implement quality in synchronous networks, but this feature is not supported on the ME 3800X and 3600X switches. We recommend configuring the SyncE network as a Resilient Ethernet Protocol (REP) segment for resiliency and to avoid timing loops when there are any network failures within the segment. See "SyncE Timing Using REP for Loop Prevention and Resiliency" section on page 6-2 Reference clocks operate in revertive or nonrevertive mode, configured by using the network-clock-select mode global configuration command. • In revertive mode, if an input clock with a higher priority than the selected reference becomes available, the higher priority reference is immediately selected. • In nonrevertive mode, if an input clock with a higher priority becomes available, the higher-priority clock is selected only when the current clock becomes invalid or unavailable. You can use the set network-clocks privileged EXEC command to configure the input reference to be either forced or automatically selected by the selection algorithm based on the highest priority valid input clock. In revertive mode, the forced clock automatically becomes the selected reference. In nonrevertive mode, the forced clock becomes the selected reference only when the existing reference is invalidated or unavailable. SyncE Timing Using REP for Loop Prevention and Resiliency In pure priority-based mode, there is a risk of timing loops if the network topology changes. You can use a REP workaround to avoid timing loops and to ensure timing resiliency. REP is a Cisco protocol used to control network loops, to respond to link failures, and to improve convergence time. See Chapter 16, "Configuring Resilient Ethernet Protocol." REP controls a group of ports connected to each other in a segment to ensure that the segment does not create any bridging loops and to respond to link failures within the segment. Cisco ME 3800X and 3600X Switch Software Configuration Guide 6-2 OL-23400-01