WO2013078997A1

WO2013078997A1 - Method for processing fault in mrp ring network, and mrp ring network

Info

Publication number: WO2013078997A1
Application number: PCT/CN2012/085454
Authority: WO
Inventors: 王兆明; 袁锦辉; 陈国华; 吴冬华; 邓桂美
Original assignee: 浙江网新技术有限公司
Priority date: 2011-11-28
Filing date: 2012-11-28
Publication date: 2013-06-06
Also published as: CN102420737A; CN102420737B

Abstract

The present invention relates to a method for processing a fault in a media redundancy protocol (MRP) ring network and an MRP ring network. The MRP ring network comprises a media redundancy management node (MRM), a plurality of media redundancy client nodes (MRC), and links for connecting the nodes. The method for processing a fault in an MRP ring network comprises: after an MRC, a link between MRCs, or a link between an MRM and an MRC is faulty, the MRM being faulty; after the MRC, the link between the MRCs, or the link between the MRM and the MRC recovers from a fault, the MRM recovering from a fault and being restarted; the MRM sending a topology change frame when being restarted. In the process that the MRM is restarted, the topology change frame is sent, so that the MRC node can perform an operation of clearing a filtering database; through the address learning and service forwarding, communication is performed again through the link recovering from the fault, thereby ensuring that the MRC can perform communication normally in the case of multiple faults.

Description

TECHNICAL FIELD The present invention relates to the field of MRP ring networks, and in particular, to a method for processing a fault in an MRP ring network and a corresponding MRP ring network. Background technique

MRP (Media Redundancy Protocol) is designed to make errors in a single switch or a single switch link in a ring network.

The MRP ring network consists of one MRM and multiple MRCs. Both MRM and MRC have the function of a switch. All kinds of Ethernet services can enter the MRP ring network through MRM or MRC connections.

As shown in FIG. 1, the MRP ring network includes: an MRM node 101, a plurality of MRC nodes 102, a terminal node 103 accessing the ring network, and a link cable 104 connecting the nodes, wherein:

The MRM node 101 is a medium redundancy management node. Each ring network needs to have one and only one MRM node under normal working state; the MRM implements the following functions:

1. Periodically send a test frame (Test Frame) to the ring network through its own two loop ports to detect the ring network state;

2. If one loop port of the MRM can receive the test frame sent by another loop port, the loop is normal. At this time, the MRM will set one of the ports (the master port, the port on which the LinkUp goes online first as the master port). To forward state 105 (FORWARDING), set the other port (secondary port) to blocking state 106 (BLOCKED) to prevent the traffic from looping;

3. If the MRM cannot receive the test frames sent by each other during the set interval, the loop has a problem. At this time, the MRM sends a Topology Change Frame from the two ports. All nodes (including MRM and MRC) in the control ring network perform the empty filtering database FDB operation after a specific time interval;

4. If the MRM is set to respond to the MRC's port change, the MRM receives the port up/down frame from the MRC.

After (LinkUp/LinkDown Frame), a topology state change frame is also issued; 5. MRM does not forward MRP protocol frames between its own two ports, including topology state change frames, test frames, and port uplink/downline frames.

The MRC node 102, which is a medium redundant client node, may have multiple MRCs in each ring network. Based on the performance requirements, the total number of nodes of the MRM and the MRC generally does not exceed 50; the MRC implements the following functions:

1. Forward the test frame and topology state change frame sent by the MRM between the two loop ports;

2. After receiving the topology state change frame, extract the time parameter carried in the topology state change frame, and perform the operation of clearing the filtering database FDB after waiting for the interval defined by the time parameter;

3. If the link status change (LinkUp/LinkDown) is detected, the port will send out the link up/down frame (LinkUp/LinkDown Frame) through two ports to notify the MRM of the change status.

The terminal node 103, the terminal node enters the ring network through the MRM or MRC connection, and is the initiator and receiver of the service; the terminal node may be a host or a common switch and a host connected thereto;

Link 104, used to connect the physical cable or logical path of each node. The link error between the two nodes indicates that the communication between the nodes is blocked, and the loop status changes accordingly.

The following describes the state machine operation and conversion relationship of MRM and MRC respectively;

The state transition of MRM is shown in Figure 2. The state machine of MRM is powered by initial state (Power_On), waiting for connection state (AC_STAT1), master port online state (PRM-UP), open loop state (CHK- R0) and closed loop state (CHK-RC);

The power-on initial state 201, complete the initialization work, including setting the two loop ports of the MRM to the BLOCKED state, and setting the related static address table;

Waiting for the connection state 202, after the initial state of power-on is completed, directly enters the waiting connection state (MRM state conversion 206), and its role is to wait for any port of the MRM to go online; in this state, since no port is online, no test frame is sent; Port online status 203, there is only one loop port online (LinkUp) state, in this state, the online port is called the main port, and is set to FORWARDING state (the other port (sub-port) is still BLOCKED Status); At the same time, start sending test frames (Test Frame) periodically to both ports (even if the secondary port is not online yet);

Open loop state 204. In this state, both loop ports are online (LinkUp) and the state is set to FORWARDING; this state is entered when there is a fault in the ring or link in the ring network (see MRM state transition 215). In the open loop state, the MRM periodically sends a test frame (Test Frame) to the two ports;

Closed loop state 205. In this state, both loop ports are in the online state (LinkUp), and the primary port is set to the FORWARDING state, and the secondary port is set to the BLOCKED state; the ring network maintains this state during normal operation; MRM periodically sends a test frame (Test Frame) to the two ports; MRM state transition 207, when in the waiting connection state 202, if there is no port uplink (LinkUp), the state machine remains in the waiting connection state 202;

The MRM state transition 208, when in the waiting connection state 202, if a port is online (LinkUp), the state machine enters the main port online state 203, and the main port is set to the FORWARDING state, and the test frame (Test Frame) is started to be sent; Conversion 209, when in the primary port online state 203, if the only uplink (LinkUp) port is again offline (LinkDown), the state machine returns to the waiting connection state 202;

The MRM state transition 210, when in the primary port online state 203, if it does not receive the test frame sent by itself and the secondary port does not go online, it remains in the current state;

The M-state transition 211, when in the open-loop state 204, if there is a port offline (LinkDown), the main port is in the online state 202; if the offline is the original primary port, the original secondary port is still Online) set to the new primary port and send the topology status change frame (because the change of the primary port will cause the change of the service direction in the loop); the MRM state transition 212, when it is in the online status of the primary port 203, if it receives itself The test frame is sent (indicating that the loop is open, NO_TC = 0 is set at this time) or the secondary port is online. At this time, both ports are online, and the state of the loop is unknown. You can continue to test through the test frame. Set NO_TC = 1) to go to closed loop state 205;

MRM state transition 213, when in the open loop state 204, if it has been unable to receive the test frame sent by itself (indicating that the loop still has a problem) or receives the port uplink/downline frame sent by the MRC (indicating that the MRC has changed, MRM Still to judge the loop state by the detection of the test frame), the state maintains the open loop state 204;

The MRM state transition 214, when in the open loop state 204, if it receives a test frame sent by itself (indicating that the loop is open, setting NO_TC = 0), at this time, the state machine goes to the closed loop state 205, and passes two Ports issue topology state change frames;

The MRM state transition 215, when in the closed loop state 205, if the test frame is not received after the interval of three consecutive test frames (the default is 1 millisecond), the state transitions to the open loop state; if the current NO_TC = 0, The topology state change frame is also sent out; if the current NO_TC = 1, the topology state change frame is not sent (see MRM state transitions 212, 214 and 217);

M-state transition 216, when in the closed-loop state 205, if there is a port offline (LinkDown), then enter the main port online state 202; if the offline is the original primary port, the original secondary port (still online ) set to the new primary port, and send the topology state change frame (because the change of the primary port will cause the change of the service direction in the loop); MRM state transition 217, when in the closed loop state 205, if it receives the test issued by itself Frame (indicating loop normal, setting NO_TC = 0), then maintaining in closed loop state 205;

The state transition of the MRC is shown in Figure 3. The state machine of the MRC is powered by the initial state (Power-On), waiting for the connection state. (AC—STATl), Data Exchange Idle, DEP, Pass Through, Data Exchange, and Pass-through Idle (PT-IDLE, Pass Through Idle) Composition; among them,

The power-on initial state 301, the initialization work is completed, including setting the two loop ports of the MRC to the BLOCKED state, and setting the related static forwarding table;

Waiting for the connection state 302, after the initial state of the power-on is completed, directly enters the waiting connection state (MRC state conversion 307), and its role is to wait for the loop port of the MRC to go online (LinkUp);

Data exchange idle state 303, this state indicates that the MRC has one and only one port is in a steady online (LinkUp) state;

Through state 304, the intermediate transition state of the MRC from one port uplink (LinkUp) to two port uplinks (LinkUp) is called the punch-through state;

Data exchange status 305, the intermediate transition state of the MRC from two port uplinks (LinkUp) to only one port uplink (LinkUp) transfer is called data exchange state;

Pass-through idle state 306, the MRC two ports are in a stable online (LinkUp) state called the punch-through idle state, at this time, both loop ports are set to FORWARDING;

MRC state transition 308, if in the waiting for connection state 302, the MRC two ports have not been online (LinkUp), the state is maintained in the waiting connection state 302;

MRC state transition 309, while waiting for connection state 302, when the MRC detects a loop port uplink (LinkUp), it is set as the master port, the port state is set to FORWARDING, enter the data exchange idle state 303;

MRC state transition 310, in the data exchange idle state 303, if the only online (LinkUp) port goes offline

(LinkDown), at this point, return to the waiting connection state 302;

The MRC state transition 311, when in the data exchange idle state 303, if only the primary port is online (LinkUp), it remains in the current state; if the topology state change frame sent by the MRM is received, the time carried in the topology state change frame is extracted. The parameter, and after waiting for the interval defined by the time parameter, the operation of clearing the filtering database FDB is performed; when the MRC state is converted 312, when the data exchange idle state 303 is reached, if the other port is also online (LinkUp), the state is converted into the punch-through state. 304, and starts to send the port uplink frame (LinkUp Frame) from the two loop ports; the MRC state transition 313, when in the punch-through state 304, starts to send the port uplink frame (see state transition 312 or state transition 318); The default is to continuously send 5 frames, one every 1 millisecond; during this time, the state remains in the punch-through state 304;

The MRC state transition 314, when in the through state 304, if a port goes offline (LinkDown), it is converted into a data exchange state 305, and stops transmitting the port uplink frame (LinkUp Frame), which will still be online (LinkUp) end The port is set as the primary port, and the port downlink frame (LinkDown Frame) is sent out from the two ports. The default is to continuously send 5 frames, and send one every 1 millisecond.

MRC state transition 315, in the through state 304 (see state transition 312 or state transition 318), if the default 5 port uplink frames are sent, the state is converted to the punch-through idle state 306; if the 5 ports are sent on the line frame In the process, after receiving the topology state change frame sent by the MRM (indicating that the MRM has observed that the MRC port state has changed and responded), the sending of the remaining port uplink frames is stopped, and the state is converted into the punch-through idle state 306;

MRC state transition 316, in data exchange state 305 (see state transition 314 or state transition 320), if the default 5 port offline frames are sent, the state is converted to data exchange idle state 303; if 5 ports are being sent During the process of the offline frame, after receiving the topology state change frame sent by the MRM (indicating that the MRM has observed that the MRC port status has changed and responded), the remaining port downlink frames are stopped, and the state is converted into data exchange. Idle state 303;

MRC state transition 317, in the data exchange state 305, if the remaining ports also go offline (LinkDown), then both ports are offline, then stop sending the port offline frame, the state is converted to wait for connection state 302;

The MRC state transition 318, when in the data exchange state 305, if the previously offline port is back online (LinkUp), then both ports are online, then the port downlink frame is stopped, the state is converted to the punch-through state 302, and the direction begins. Outgoing port uplink frame (LinkUp Frame);

The MRC state transition 319, when in the data exchange state 305, begins to send the port offline frame (see state transition 314 or state transition 320); the default is to continuously send 5 frames, one every 1 millisecond; , the state remains in the data exchange state 305;

The MRC state is converted to 320. When the through state is idle 306, if a port goes offline (LinkDown), the state is converted into a data exchange state 305, and the port that is still online (LinkUp) is set as the master port, and starts to go out. Send port offline frame;

The MRC state transition 321 is in the through state idle state 306. If the port offline (LinkDown) does not occur, the state remains in this state; now, a link in the ring network fails, and then the link is restored. The specific process of MRP ring network redundancy protection:

Under normal working conditions, the state of the MRP ring network is as shown in Figure 4, where

BL0CKED404, indicating that the secondary port of the MRM is set to the BLOCKED state to prevent the service from being looped;

Dotted line 407, although the link 407 is physically connected, since the secondary port of the MRM is set to the BLOCKED state, the service does not pass the link indicated by 407 (the solid line represents the path of the service between the nodes, and the dotted line represents the link due to the failure or port BLOCKED can not pass the business, the same below);

MRC 0, as shown by 401 in Figure 4, a normal MRC node, assuming that there is traffic communicating with MRC 1 402;

MRC 1, as shown by 402 in Figure 4, a normal MRC node, assuming that there is traffic communicating with MRC 0 401;

Link 403, in the normal state, MRC 0 401 and MRC 1 402 communicate through the directly connected link 403 (because the MRM sub-port is set to the BLOCKED state, MRC 0 and MRC 1 pass the address learning to select the link 403 To communicate); If a link between MRCs fails at a certain time, as shown in Figure 5, according to the protocol state machine conversion process, the following steps occur:

Step 401, the ring network is in a normal state, the MRM is in a closed loop state 205, and according to the state transition 217, at this time NO_TC = 0;

Step 402, at a certain moment, the link between the MRC 0 501 and the MRC 1 502 in FIG. 5 fails (link 503 is disconnected); Step 403, the two ports of the MRM cannot receive themselves due to the failure of the intermediate link The sent test frame is converted into 215 by the MRM state, and the MRM enters the open loop state 204, and both ports are set to the FORWARDING state, and the topology state change frame is sent out;

Step 404: After receiving the topology state change frame, each node in the ring network changes the time interval parameter in the frame according to the topology state, starts a countdown, and performs an operation of clearing the filtering database FDB after the countdown is completed;

Step 405: After all the nodes perform the operation of clearing and filtering the database FDB, the node performs address learning and service forwarding operations again, and the service is turned on again;

In step 406, the loop state transition is as shown in FIG. 5, wherein the dotted line represents that no traffic passes due to the link 503 failure, and the traffic circulates through the path (links 506, 507, and 508) indicated by the solid line. If the link failure is fixed at a certain time, according to the protocol state machine conversion process, the following steps occur:

Step 501, the link failure point is restored, and the link between the MRC 0 501 and the MRC 1 502 is reconnected normally; Step 502, since the entire link of the ring network is normal, the two loop ports of the MRM are re-received and mutually sent out. Test frame (Test Frame);

Step 503, by MRM state conversion 214, the MRM enters a closed loop state 205, sets the secondary port from FORWARDING to BLOCKED, and issues a topology state change frame from the two loop ports;

Step 504: After receiving the topology state change frame, each node in the ring network changes the time interval parameter in the frame according to the topology state, starts a countdown, and performs an operation of clearing the filtering database FDB after the countdown is completed;

Step 505: After all the nodes perform the operation of clearing the filtering database FDB, the address learning and the service forwarding are performed again. Operation, the business is turned on again;

Step 506, the node loop state is restored to the state shown in FIG. 4, wherein the dotted line represents that the link has no traffic due to the MRM secondary port being in the BLOCKED state, and the traffic is circulated through the line (link 403) indicated by the solid line;

Steps 401-406 and steps 501-506 respectively represent the process of protecting the traffic flow according to the MRP protocol after a link error in the ring network and an incorrect link recovery.

Although the above process only shows the process of recovering after a link failure between MRCs, the link failure between MRM and MRC and the recovery process are the same as those described above, and will not be described again. The following is a list of situations that may occur in reality (the link between the MRC or the MRC node fails simultaneously with the MRM, and both recover) and the service communication between the nodes cannot be restored using only the existing protocol:

Step 601: In the ring state diagram of the normal operation shown in FIG. 4, the MRC 0 401 and the MRC 1 402 communicate, and in the normal state, the MRC 0 401 and the MRC 1 402 are connected by the link 403 directly connected to the two. Communication; Step 602, at some point, the link between MRC 0 401 and MRC 1 402 fails (see the link in Figure 5)

503), after the MRM detects the fault (cannot receive the test frame sent by itself), issues the topology state change frame (Topology

Change Frame), after all the nodes clear the filtering database FDB, the service re-addresses the learning, after which MRC 0 and MRC 1 communicate through the link through the MRM node, that is, step 401-step 406;

Step 603, assuming that the MRM is restarted due to the failure reason (shown by the MRM dotted line 605 in FIG. 6), during the MRM restart, the previously failed link 503 (ie, the link 603 in FIG. 6) is also repaired. , reclose, as shown in Figure 6;

Step 604, after the MRM restart is completed, because the fault in the loop has been eliminated, the state machine of the MRM is initialized by power-on 201-> waiting for the connection state 202-> (state transition 208) the master port online state 203-> (state transition 212, The condition is that the secondary port is online, and NO_TC = 1 is set. The closed loop state 205 is used for state transition. Since the two ports of the MRM can receive the test frame sent by themselves, according to the state transition 217, the MRM is maintained in the closed loop state. 205, and its secondary port is in a BLOCKED state;

Step 605, since the MDB 0 and the FDB inside the MRC 1 do not perform the clear operation (because NO_TC = 1, the MRM does not issue the topology state change frame to let each node clear the filtering database FDB), and still retains the service forwarding path previously through the MRM. (ie, links 706, 707, and 708 in FIG. 7), and the secondary port of the MRM is already in the BLOCKED 704 state, and the service cannot pass (shown by the dotted link 707); therefore, although the physical connection of the loop has been restored, the MRC 0 still cannot communicate with MRC 1, as shown in Figure 7.

Although the above process only shows the problems occurring in the recovery process after the MRC link failure and the MRM failure occur simultaneously, the link failure between the MRM and the MRC and the MRM failure occur simultaneously after the recovery process and the above Problem The same problem will not be repeated. SUMMARY OF THE INVENTION The present invention provides a method for handling a fault in a medium redundancy protocol MRP ring network, which ensures that the MRC can communicate normally when an MRC fault and an MRM fault occur simultaneously, or when a link fault and an MRM fault occur simultaneously.

The present invention also provides an MRP ring network in which normal ring communication can be ensured if an MRC failure or a link failure and an MRM failure occur simultaneously.

In order to achieve the above object, the technical solution of the present invention is implemented as follows:

A method for processing a fault in a medium redundancy protocol MRP ring network, the MRP ring network comprising a medium redundancy management node MRM, a plurality of medium redundancy client nodes MRC, and a link connecting each node, wherein The method includes the following steps: after the MRC, or the link between the MRCs, or the link between the MRM and the MRC fails, the MRM also fails;

After the failure of the link between the MRC or the MRC is recovered, the failure of the MRM is restored and restarted;

The MRM sends a topology state change frame during the restart.

A medium redundancy protocol MRP ring network, the MRP ring network includes a medium redundancy management node MRM, a plurality of medium redundant client nodes MRC and a link, wherein:

The MRC, connecting the other MRC or the MRM through the link;

The link is used to connect the MRM and the multiple MRCs into the MRP ring network;

The MRM is configured to manage the MRP ring network, after the MRC, or the link between the MRCs, or the link between the MRM and the MRC fails, if the MRM also fails And after the MRC, or the link between the MRCs, or the failure of the link between the MRM and the MRC, the failure of the MRM is also restored and restarted, then the MRM A topology state change frame is sent during the restart.

The method for processing a fault in the MRP ring network and the MRP ring network provided by the present invention, by sending a topology state change frame during the MRM restart process, enables the MRC node to perform the operation of clearing the filtering database FDB. Therefore, the MRM service forwarding path originally reserved by the MRC is cleared, and the address recovery and the service forwarding are used to re-transmit the recovered faulty link to ensure that the MRC can communicate normally in the event of multiple faults. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram of an MRP ring network;

2 is a schematic diagram of state transition of an MRM; Figure 3 is a schematic diagram showing the state transition of the MRC;

Figure 4 is a state diagram of the ring network during normal operation;

Figure 5 is a state diagram of the ring network after a link failure occurs;

Figure 6 shows the MRM restart operation and the faulty link recovery;

Figure 7 is a schematic diagram of the communication state between MRCs after the MRM restart is completed. DETAILED DESCRIPTION The following terms are explained:

IEC : International Electrotechnical Commission, International Electrotechnical Commissioner;

MRP: Media Redundancy Protocol, media redundancy protocol, defined by IEC 62439-2. Originally developed by Siemens, a redundant protection protocol based on a ring structure;

MRM: Media Redundancy Manager, media redundancy management node, MRP administrator role defined by MRP protocol, only one MRM is allowed in each MRP ring network;

NO— TC : No Topology Change, suppresses topology state change; NO—TC is an indicator variable inside MRM. When NO—TC=1, no topology state change frame is sent. When NO—TC=0, the topology can be sent. State change frame;

MRC: Media Redundancy Client, media redundant client node, MRP client role defined by MRP protocol, one or more MRCs are allowed in each MRP ring network;

Ring Port: The two ports of the MRM or MRC connected ring network are called loop ports. Each node has two loop ports connected to the same ring network. Among the two ring ports, first The port of the LinkUp is called the Primary Ring Port, and the other is called the Secondary Ring Port.

FDB: Filtering Data Base, filtering database, lookup table established and maintained by the Ethernet switch for forwarding data frames, also called Forwarding Table;

Static forwarding table: The forwarding entry for the special address configured by the administrator is configured according to the application requirements. Clearing the filtering database (FDB): The action of the Ethernet switch to clear the internal filtering database (FDB). Affected;

Address learning: The Ethernet switch records the source address of the Ethernet frame entering a port and the entered port number into the filtering database for subsequent address lookup;

Service forwarding: The Ethernet switch searches for the corresponding entry from the filtering database according to the destination address of the received Ethernet frame. If there is an entry corresponding to the destination address in the database, the frame is forwarded from the port specified in the entry. Go out; if the corresponding entry is not found, broadcast it from all ports, also called FLOODING operation;

Test Frame: MRP test frame, MRM is used to detect the ring network status and periodically send specific MRP protocol frames; Topology Change Frame: The MRP control frame sent by all the nodes of the ring network after the MRM detects the change of the state of the ring network. This frame contains a time parameter for clearing the filtering database FDB (default is 1. 5- 0. 5 milliseconds), indicating how many time intervals the node has to empty the filtering database FDB; after issuing such a frame, the MRM starts the countdown according to the time parameter carried in the sending frame, and after the countdown is completed, clears the filtering database. FDB operation; after receiving such a frame, each MRC node extracts the time parameter carried in the frame, and starts the countdown, and after the countdown is completed, performs the operation of clearing the filtering database FDB;

LinkUp I LinkDown Frame: The MRP port uplink/downline message frame is sent by the MRC. After the MRC detects that the loop port of the local node is LinkUp/LinkDown, this message frame is sent to notify the MRM to take the corresponding action.

FORWARDING: Forward state, one of the MRP loop port states, in which all frames are forwarded. In the case of the MRM restart process, if the original fault link (including the fault between the MRC and the fault between the MRM and the MRC) or the fault MRC is also restored, the inter-node communication may still be blocked, and the MRM in the existing MRP protocol The state machine has been modified. The specific method is:

Step 701: When the MRM receives a port uplink (LinkUp) signal, and the state machine is converted from the waiting connection state 202 to the main port online state 203, the original state machine action only sets the main port to the FORWARDING state, and starts to send the test. A test frame, as shown in the state transition 208 in FIG. 2; the method proposed by the present invention is: adding a topology change frame (Topology Change Frame) behavior based on the original behavior (state transition 208);

Step 702: Perform a restart operation on the faulty MRM in FIG. 6 according to the method in step 701, and recover the fault link or the fault MRC in the ring network. When the MRM is in the restart process, if the faulty link has been Recovery, when the MRM detects that a loop port (primary port) is online, it is converted 208 by the state in FIG. 2, and the state machine transits from the waiting connection state 202 to the main port online state 203, and simultaneously issues a topology state change frame; After receiving the topology state change frame sent by the MRM in step 702, each node of the ring network extracts the time parameter carried in the frame, and starts the countdown. After the countdown is completed, the FDB operation is performed to clear the filter database; As shown in FIG. 7, the MRM service forwarding path originally reserved by the MRC 0 701 and the MRC 1 702 is cleared, and the two communicate through the link 703 through address learning and service forwarding, that is, the service is resumed;

Step 704, in synchronization with step 703, when the state machine of the MRM detects that the second port (the secondary port) is also online, the state transition 212 from the state in FIG. 2, and the MRM is changed from the primary port online state 203 to the closed loop state 205. And the secondary port is set to the BLOCKED state; since the loop is physically looped at this time, since the two ports of the MRM can receive the test frame sent by themselves, according to the state transition 217, the MRM is maintained in the closed loop state 205; Step 705, the state of the ring network after completing step 701 704 is as shown in FIG. 4.

Since the original MRP protocol can only handle one node failure or link failure, in actual applications, there may be two or more failure points (including MRM), when the link between MRC, or MRC, or After the link between the MRC and the MRM and the MRM fail simultaneously, if the MRM is restored (or restarted) after the other failure points are recovered, if only the original protocol is used, the service communication cannot be recovered due to the topology change ( Steps 601-605); Therefore, when the MRM transitions from the waiting connection state 202 to the primary port online state 203, the topology change frame is added to re-switch the service (steps 701-705);

Based on the above method:

1. Under normal circumstances, during the MRM startup process, if a loop port is online (state transition 208 in Figure 2), a topology state change frame will also be issued, and all nodes will simultaneously clear the filter database FDB; State, node emptying the filtering database FDB does not cause problems;

2. In the case of multiple nodes (including MRM) or link failures and recovery in succession:

If the MRM is restarted after the other nodes are restarted, the topology change state frame is sent, and the previously started node is emptied to filter the database FDB operation, thereby performing address learning and service forwarding, so that the service re-establishes communication;

If the MRM is restarted before other nodes, the MRM will detect the loop status by periodically sending a test frame after the MRM works normally. If the ring network still has a fault, the MRM will send a topology state change frame according to the state transition 215, so that the service is enabled. Relearning and restoring; if the fault is recovered, according to the state transition 214, the MRM will also issue a topology state change frame to re-learn and recover the service.

The above description is only the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modifications, equivalents, improvements, etc., which are included in the spirit and principles of the present invention, should be included in Within the scope of protection of the present invention.

Claims

Claim

A method for processing a fault in a medium redundancy protocol MRP ring network, the MRP ring network comprising a medium redundancy management node MRM, a plurality of medium redundant client nodes MRC, and a link connecting the nodes, the characteristics thereof It consists of the following steps:

After the MRC, or the link between the MRCs, or the link between the MRM and the MRC fails, the MRM also fails;

After the MRC, or the link between the MRCs, or the failure of the link between the MRM and the MRC, the failure of the MRM is restored and restarted;

The MRM sends a topology state change frame during the restart.

The method according to claim 1, wherein in the restarting of the MRM, a topology state change frame is sent when a state of the MRM transitions from a waiting connection state to a primary port online state.

The method according to claim 2, wherein, after the MRM detects that a loop port of the MRM is online, setting the loop port of the uplink to a forwarding FORWARDING state, the MRM The state transitions from waiting for the connection state to the primary port online state.

The method according to claim 3, wherein the test frame is further sent when the state of the MRM transitions from the waiting connection state to the main port online state.

The method according to any one of claims 1 to 4, wherein the method further comprises: after the MRM sends a topology state change frame, receiving an MRC of the topology state change frame to clear the filter database FDB. Operation, address learning and service forwarding, so that the service re-establishes communication.

The method of claim 5, wherein the operation of the MRC that receives the topology state change frame to perform the clearing of the filtering database FDB comprises: receiving the topology state change frame of the MRC to extract the topology state change frame carried The time parameter is, and waits for the time interval defined by the time parameter to perform the operation of clearing the filtering database FDB.

The method according to claim 5, wherein, by clearing the operation of filtering the database FDB, the MRM-preserved service forwarding path reserved by the MRC is cleared, and the link between the MRC or the MRC after the fault recovery is resumed. Communication.

The method according to any one of claims 3-4, wherein: after the MRM detects that another loop port of the MRM is also online, the MRM is changed from a primary port online state to a closed loop state. And set another loop port of the upper line to block the BLOCKED state.

The method according to claim 1, wherein the MRM discovers that the MRC, or a link between the MRCs, or a link between the MRM and the MRC occurs by sending a test frame. malfunction.

The method according to claim 9, wherein the MRM discovers the MRC by sending a test frame, Or the link between the MRC or the link between the MRM and the MRC specifically includes: when the MRM cannot receive the test frame sent by itself, the MRC or the chain between the MRCs is found. The road, or the link between the MRM and the MRC fails.

The method according to any one of claims 9 to 10, wherein after the MRC, or the link between the MRCs, or the link between the MRM and the MRC fails, Before the MRM fails, the state machine of the MRM enters an open loop state, and the MRM sets the two ports to forward the FORWARDING state, and sends the topology state change frame to the outside.

The method according to claim 11, wherein after the MRC, or a link between the MRCs, or a link between the MRM and the MRC fails, and the MRM fails Previously, the MRC node receives the topology state change frame, performs an empty filter database FDB operation, performs address learning again, and communicates through the link through the MRM.

A media redundancy protocol MRP ring network, the MRP ring network comprising a media redundancy management node MRM, a plurality of media redundant client nodes MRC and a link, wherein:

The MRC, connecting the other MRC or the MRM through the link;

The MRP ring network according to claim 13, wherein, in the process of restarting the MRM, when the state of the MRM is changed from a waiting connection state to a main port online state, a topology state change frame is sent. .

The MRP ring network according to claim 14, wherein, after the MRM detects that a loop port is online, the loop port of the uplink is set to forward a FORWARDING state, and the state of the MRM is from Waiting for the connection state to transition to the primary port online state.

The MRP ring network according to claim 15, wherein a test frame is further transmitted when a state of the MRM transitions from a waiting connection state to a master port online state.

The MRP ring network according to any one of claims 13 to 16, wherein the MRC is further configured to perform an operation of clearing the filtering database FDB after receiving the topology state change frame sent by the MRM, and performing Address learning and service forwarding enable the service to re-establish communication.

The MRP ring network according to claim 17, wherein the operation of the MRC to clear the filtering database FDB comprises: extracting, by the MRC, a time parameter carried in the topology state change frame, and waiting for the time defined by the time parameter Room After the interval, the operation of filtering the database FDB is emptied.

The MRP ring network according to claim 17, wherein the operation forwarding path of the MRM retained by the MRC is cleared by the operation of clearing the filtering database FDB, and the MRC or the MRC between the fault recovery is re-passed. The link communicates.

The MRP ring network according to any one of claims 15-16, wherein the MRM is further configured to: after detecting that another loop port of the MRM is also online, the online status of the main port is changed to The closed loop state, and the other loop port of the upper line is set to block the BLOCKED state.

The MRP ring network according to claim 13, wherein the MRM is further configured to send a test frame to discover the MRC, or a link between the MRCs, or between the MRM and the MRC. The link has failed.

The MRP ring network according to claim 21, wherein the MRM sends a test frame to discover the link between the MRC, or the MRC, or a link between the MRM and the MRC. The failure includes: when the MRM cannot receive the test frame sent by itself, the link between the MRC, or the MRC, or the link between the MRM and the MRC is found to be faulty.

The method according to any one of claims 21 to 22, wherein after the MRC, or the link between the MRCs, or the link between the MRM and the MRC fails, Before the MRM fails, the state of the MRM enters an open loop state, and the MRM sets the two ports to forward the FORWARDING state, and sends the topology state change frame to the outside.

The MRP ring network according to claim 23, wherein after the MRC, or a link between the MRCs, or a link between the MRM and the MRC fails, and the MRM Before the fault occurs, the MRC node is further configured to perform an operation of clearing the filtering database FDB after receiving the topology state change frame, and performing address learning again, by communicating through the link of the MRM.

The MRP ring network according to any one of claims 13-16 and 21-22, wherein the MRP ring network further comprises a terminal node that accesses the ring network through the MRM or the MRC.