RLDP (Ring Light Detection and Protection) of Simplified Optical 3.0--Function Introduction

The RLDP (Rapid Link Detection Protocol) is an Ethernet link fault detection protocol.  
Generally, the Ethernet link detection mechanism only uses the physical connection status to detect link connectivity through automatic negotiation at the physical layer.
However, this detection mechanism has certain limitations. In some cases, it cannot provide users with reliable link detection information. For example, the optical fiber receiving line on the optical fiber port is connected incorrectly. Due to the existence of optical fiber converter, the corresponding port of the device is physically linkup, but the actual layer 2 link cannot communicate. Another example is that an intermediate network is set up between two Ethernet devices. Due to the existence of network transmission relay devices, if these relay devices fail, the same problem will be caused, Using RLDP, you can easily and quickly detect the link faults of Ethernet devices, including unidirectional link faults, bidirectional link faults, and loop link faults. If a fault is detected, RLDP will automatically close the relevant ports according to the user's configured fault handling method or notify the user to manually close them to avoid incorrect forwarding of traffic or prevent Ethernet layer 2 loops.
1.1 Basic Concepts
1.1.1 link one-way failure
One-way link failure refers to the end of the optical fiber transmission link being disconnected or not connected, or one of the lines in the twisted pair being disconnected, or the intermediate device between two equipment having a one-way disconnect, which will cause traffic to be wrongly forwarded or the loop protection protocol (such as STP) to fail.
1.1.2 link bidirectional fault
Link bidirectional fault refers to the situation where two optical fibers are broken, or two lines in the twisted pair are broken, or there is bidirectional interruption between two devices, which will result in traffic being incorrectly forwarded.
1.1.3 link loop failure
The link loop fault refers to the fault caused by the incorrect connection of devices downstream to other devices, forming a loop, which will cause a broadcast storm.
1.1.4 Protocol messages
The RLDP protocol defines the following three types of messages:
Probe message (Prob): a Layer 2 multicast message for neighbor negotiation, unidirectional or bidirectional link detection, the default encapsulation format of the message is of SNAP type, and the encapsulation method is automatically changed to Ethernet II if the neighbor sends out a message in Ethernet II format.
Echo message (Echo): Layer 2 unicast message in response to Prob message, used for unidirectional or bidirectional link detection, the default encapsulation format of the message is SNAP, and the encapsulation method is automatically changed to Ethernet II if the neighbor sends out a message in Ethernet II format.
Loop: Layer 2 multicast message for downlink loop detection, this type of message will only be received by the sender, the encapsulation format of the message is SNAP encapsulation.
1.1.5 Probe Interval
RLDP can configure the probing interval and the maximum number of probes. The probe interval determines the sending period of Prob and Loop messages, and the device responds to an Echo message immediately after receiving a Prob message. The Probe Interval and Maximum Probe Count determine the maximum probe time (Probe Interval × Maximum Probe Count + 1) for unidirectional or bidirectional link probing, and the unidirectional or bidirectional fault processing can be triggered if the neighbor's Prob message or Echo message cannot be received correctly within the maximum probe time. The timers involved are described in Table 1-1.

Timer Name	Timer description
Send Timer	After the link detection function is enabled, the sending timer is enabled to periodically send Prob packets or Loop packets. The interval for sending the timer is the configured detection interval
Timeout Timer	● When a unidirectional or bidirectional link fault is detected, a timeout timer is started after the Prob packet is sent. The timeout period is the maximum detection time (detection interval x maximum detection times + 1). ● In unidirectional link detection, when the timeout period expires and the port receives only probe packets or reply packets from the neighbor, a unidirectional link fault occurs on the link ● In bidirectional link detection, if the port cannot receive neither probe packets nor reply packets from the neighbor after the timeout period expires, a bidirectional link fault occurs on the link

1.1.6 Neighbor Negotiation
A port configured with unidirectional or bidirectional detection can learn from the peer device as a neighbor. One port can learn one neighbor. The neighbor can change. After the negotiation function is enabled, unidirectional or bidirectional detection is performed only after the port is negotiated with the neighbor. If the port successfully receives a Prob packet from the neighbor during the negotiation, the negotiation is considered successful. However, if the protocol is started only when the fault exists, the detection cannot be started because the neighbor cannot be learned. Therefore, you are advised to restore the link error state first.
1.1.7 Associated Action
Table 1-2 lists the methods of associating and recovering an RLDP port after a fault occurs.

Type	Associated action/recovery mode	Description
Associated action after a fault occurs	Warning	Only related syslogs are printed to describe the current fault port and fault type
	Shutdown SVI	After Syslog is printed, if the faulty port is a physical switch port or L2 AP member port, the system queries the SVI based on the Access VLAN or Native VLAN to which the port belongs and runs the Shutdown operation.
	Port violation	After Syslog is printed, the faulty port is set to the violation state. In this case, the port physically enters the Link down state.
	Block	After Syslog is printed, the forwarding status of the faulty port is set to Block. In this case, the port does not forward the received packets.
Fault recovery mode	Manually Reset	All faulty ports are manually restored to the initial state. In this case, link detection is restarted.
	Manual or automatic execution Errdisable Recovery	Manually or periodically (the default value is 30 seconds and can be configured) restore all faulty ports to the initialization state and restart link detection.
	Automatic recovery	In the case of unidirectional or bidirectional link detection, if the specified troubleshooting mode is not a port violation, you can rely on the Prob packets exchanged with the neighbor to automatically restore the initial state and restart the link detection.
	Manual or automatic execution Errdisable Recovery	Periodically (configurable) restore all failed ports to the initial state and restart link detection.

Table 1-2

1.1.8 Status of the RLDP port
The status of an RLDP port can be as follows:
Normal: indicates the status of the port after the detection is enabled.
Error: indicates the status of a link fault detected on a port. The fault can be caused by a unidirectional, bidirectional, or loop fault
1.1.9 Detection principle
Generally, the Ethernet link detection mechanism only uses the physical connection status to detect link connectivity through automatic negotiation at the physical layer. However, this detection mechanism has certain limitations. In some cases, although the physical layer is connected and can work normally, the actual layer 2 link cannot communicate or is abnormal. RLDP uses detection packets, response packets, or loop packets to identify neighbor devices and detect link faults.
Figure 1-3 shows the RLDP detection logic.

Figure 1-3

Table 1-4 lists the four main states.

State	Explain
DISABLE	After the function of an RLDP port is enabled, the port is in the DISABLE state. Unidirectional link detection, bidirectional link detection, and downlink loop detection are not enabled on the port. If the configuration is disabled in any state, the port is in the disabled state.
INIT	The initial status of the RLDP protocol after it is enabled or after the fault is recovered is configured. In this case, unidirectional link detection, bidirectional link detection, or downlink loop detection is enabled on the port, and the corresponding processing mode is specified when a link fault occurs, or the RLDP recovers from a unidirectional fault, bidirectional fault, or loop fault.
NORMAL	RLDP enters the status after the detection is enabled. RLDP enters the status after the detection of downstream loops is enabled. RLDP enters the status after the detection of unidirectional or bidirectional links is enabled.
ERROR	RLDP detects a fault. The port detects a unidirectional link fault or a bidirectional link fault. When a fault occurs, handle it according to the fault handling mode specified by the corresponding detection mode.

Table 1-4

1.1.10 Establishing an RLDP detection.
The link detection modes of RLDP mainly include unidirectional link detection, bidirectional link detection, and downlink loop detection.
1.1.11 RLDP Unidirectional Link Detection
After unidirectional link detection is enabled, the port periodically sends Prob packets, receives Echo packets from the neighbor, receives Prob packets from the neighbor, and responds Echo packets to the neighbor in time. Within the maximum detection time, if you can receive only the Prob packets of the neighbor but not the Echo packets of the neighbor or neither the Prob packets nor the Echo packets of the neighbor, the unidirectional fault processing is triggered and the detection stops.
1.1.12 RLDP Bidirectional Link Detection
After bidirectional link detection is enabled, the port periodically sends Prob packets, receives Echo packets from the neighbor, receives Prob packets from the neighbor, and responds Echo packets to the neighbor in time. Within the maximum detection time, if neither Prob nor Echo packets are received from the neighbor, bidirectional fault processing is triggered and the detection stops.
1.1.13 Detecting the RLDP Link loop
After the down-link Loop detection is enabled, the ports periodically send Loop packets. After receiving LOOP packets, if the sending port and the receiving port are routing ports or L3 AP member ports, and the sending port and the receiving port are the same, the loop fault is triggered. If the sending and receiving ports are switching ports or L2 AP member ports, the default vlans of the ports are the same, and the forwarding status of both ports is Forward, the loop fault is triggered. After the fault occurs, the fault is handled according to the corresponding troubleshooting method and the detection stops.