CN102088493B

CN102088493B - Selective address refreshing method and system of Ethernet multi-ring network

Info

Publication number: CN102088493B
Application number: CN2009102255894A
Authority: CN
Inventors: 王斌; 江丽莉
Original assignee: ZTE Corp
Current assignee: ZTE Corp
Priority date: 2009-12-03
Filing date: 2009-12-03
Publication date: 2013-11-06
Anticipated expiration: 2029-12-03
Also published as: CN102088493A

Abstract

The invention discloses a selective address refreshing method of an Ethernet multi-ring network. The method comprises the following steps of: realizing selective address refreshing by using the Ethernet multi-ring network through a loopback data message; and only transmitting the loopback data message which enters a loop on the loop. The invention also discloses a selective address refreshing system of the Ethernet multi-ring network. In the system, a refreshing unit is used for realizing selective address refreshing by using the Ethernet multi-ring network through the loopback data message; and a transmission unit is used for transmitting the loopback data message which enters the loop on the loop. Due to the adoption of the method and the system, the performance of the Ethernet multi-ring network can be promoted.

Description

Selective address refreshing method and system of Ethernet multi-ring network

Technical Field

The invention relates to the field of data communication, in particular to a selective address refreshing method and system of an Ethernet multi-ring network.

Background

In the practical application of the ethernet, various protection technologies are widely adopted to implement the redundant backup between the primary path and the backup path. When the main path and the standby path are intact, the protection data forwarding function of the standby path is blocked, and the protection data between networks is transmitted on the main path; when the main path fails, the protection data forwarding function of the standby path is opened, and the protection data between networks is switched to the standby path for transmission, so that the protection data is prevented from being repeatedly received and forming broadcast storm under the normal state of the networks, the standby path is started to transmit the protection data when the main path of the networks fails, the fault resistance of the Ethernet is improved, and the requirement of high real-time performance that the convergence time is less than 50ms during switching is met. The protection technology of the existing ethernet multi-ring network is illustrated below.

For example, as shown in fig. 1, which is a schematic structural diagram of an ethernet multi-ring network, in fig. 1, nodes S1 to S6 are ethernet switches, and network B is connected to node S2, and network a is connected to node S5. Communication is performed between network a and network B. There are four physical paths between network a and network B, namely: network a ← → node S5 ← → node S3 ← → node S2 ← → network B; network a ← → node S5 ← → node S3 ← → node S4 ← → node S1 ← → node S2 ← → network B; network a ← → node S5 ← → node S6 ← → node S4 ← → node S3 ← → node S2 ← → network B; network a ← → node S5 ← → node S6 ← → node S4 ← → node S1 ← → node S2 ← → network B.

When the ethernet multi-ring protection technology is applied, a ring and a sub-ring are generally adopted. Wherein, namely, the ring is a complete Ethernet ring; a sub-ring is an ethernet ring that is connected to other rings or networks through interconnecting nodes. An interconnect node is a common node belonging to two or more ethernet rings simultaneously, and may also be referred to as a shared node. As shown in FIG. 2, which contains a Ring and a sub-Ring, Ring1 is the Ring and Ring2 is the sub-Ring. Ring1 contains nodes S1, S2, S3 and S4, and contains links: < S1, S2>, < S2, S3>, < S3, S4> and < S4, S1 >; ring2 comprises nodes S3, S5, S6 and S4, and comprises links: < S3, S5>, < S5, S6> and < S6, S4 >. It is particularly emphasized that the < S3, S4> links belong to Ring1 and not Ring 2. Wherein the nodes S3 and S4 are interconnected nodes of Ring1 and Ring2, and the 33 port of the node S3 and the 43 port of the node S4 belong to Ring2, called access ports. In a ring network, when the ring network has no fault, a link needs to be in a blocking state for forwarding a data packet in one ring to prevent looping, the link is generally called a ring protection link or a normally blocked link, and the ring protection link participates in switching between an active path and a protection path in the ring. The node that owns the ring protection link is referred to herein as the ring protection link control node. As shown in fig. 3, in Ring1, the node S4 is a Ring protection link control node, and the direct link with the port 41 of the node S4 is a Ring protection link of Ring 1. In Ring2, the node S5 is a Ring protection link control node, and the direct link with the port 52 of the node S5 is a Ring protection link of Ring 2. Under normal conditions, Ring protection link control nodes of Ring1 and Ring2 block the data forwarding function of their slave ports, preventing protection data from being repeatedly forwarded and forming broadcast storms.

When the links of the Ethernet multi-ring network are all intact, the ring protection link control nodes of the ring and the subring block the protection data forwarding function of the slave port. As shown in fig. 2, the node S4 blocks the protection data forwarding function of the port 41, the node S5 blocks the protection data forwarding function of the port 52, and the communication paths of the networks B and a are: network B ← → node S2 ← → S3 ← → S5 ← → network a.

When the link of the Ethernet multi-ring network fails, if the failed link is not a protection link, the ring protection link control node opens the protection data forwarding function of the slave port, and each node also needs to refresh an address forwarding table, and the communication between the networks is transmitted according to a new path. As shown in fig. 3, when a link < S2, S3> on Ring1 fails, and when detecting a link failure, node S2 and node S3 respectively block the data forwarding function of port 22 and port 31 to notify other nodes that a link fails, and when receiving the failure notification, node S4 (Ring protection link control node) opens the protection data forwarding function of port 41, and each node on Ring1 also refreshes the address forwarding table, the new communication paths of networks B and a are: network B ← → node S2 ← → node S1 ← → node S4 ← → node S3 ← → node S5 ← → network a.

When the link of the ethernet multi-ring network is recovered, recovery switching is performed, the transmission path when the network transmission is recovered to a normal state, and due to the path change, the node also needs to refresh the address forwarding table.

In addition, if the nodes of the sub-ring refresh addresses, but the nodes of the ring do not refresh addresses, the following problems may be caused.

For example, in fig. 4, when the ethernet multiple ring network is failure-free, the communication paths of networks B and a are: network B ← → node S2 ← → S3 ← → S5 ← → network a. When the link of the sub-Ring 2 fails, the link < S3, S5> of Ring2 fails, and the node S5 detects that the link corresponding to the port 51 fails, refreshes the address forwarding table, blocks the protection data forwarding function of the port 51, opens the protection data forwarding function of the port 52, and sends a failure status frame to the outside. After receiving the failure status frame, the node S6 refreshes the address forwarding table, and a new transmission path is formed between the network a and the network B. After the node S5 and the node S6 refresh the address forwarding table, the protection data sent by the network a to the network B can finally reach the network B through the broadcast of the node S5 and the node S6, and meanwhile, each node learns the address of the network a. However, before network a sends network B data, if network B sends data to network a, a large amount of packet loss occurs. This is because the node S2 has not refreshed the address forwarding table, the address forwarding table is also an entry before path switching, i.e. an incorrect address entry, and the protection data sent by the network B to the network a is still forwarded according to the incorrect address forwarding table, i.e. sent from the egress port 22 of the node S2, and these data cannot actually reach the network a due to link failure and port blocking, and can only reach the network B after waiting for the switch to learn the correct address egress port of the network a, so the path switching time of the network B to the network a depends on whether there is traffic sent by the network a to the network B, and this time sometimes even exceeds 50 ms. Similar problems exist with communication between networks a and B when the ethernet ring performs a protection switch for recovery.

Currently, in ITU-T g.8032, in order to solve the problem of an error occurring in an entry of a MAC address table due to link switching, a protocol packet carrying address refresh information is usually used to complete the process.

As shown in fig. 5, Ring1 is a complete closed Ring, and includes nodes such as S1, S2, S3, S4, S5, and S6, and links therebetween, where the node S2 is a control node, and it is a forwarding function that blocks data packets of 21 ports and Ring2 protocol packets (the Ring2 protocol packet is treated as a data packet in the Ring 1) under the condition that Ring1 has no fault; ring2 is a sub-Ring, comprising S1, S8, S7 and S3 nodes, and S7 node is a control node, which blocks the data packet forwarding function of 72 port without fault in Ring2, but does not block the protocol packet on the control Virtual Local Area Network (VLAN) of Ring 2. In the case of ethernet multiple ring network without failure, the communication paths of network a and network B are:

network a ← → node S4 ← → S5 ← → S6 ← → S1 ← → S8 ← → network B.

As shown in fig. 6, when a link < S1, S8> of Ring2 has failed, and a node S1 and S8 detect the failure, respectively refresh their address forwarding tables and block the ports adjacent to the failed link, and at the same time, the S8 node also periodically sends an SF protocol message, i.e., a failure notification message, along port 82, refreshes their address forwarding tables after the first receipt of the SF message by other nodes of Ring2, and after the first receipt of the SF message and the refreshing of the address forwarding tables by the node S3, i.e., the interconnection node, generates a new Flush message on the control VLAN of Ring1 and sends it continuously 3 times. And after receiving the Flush message, the nodes on the Ring1 refresh respective address forwarding tables. When the nodes on the ethernet multi-ring network learn the MAC addresses of the network a and the network B, the communication path between the network a and the network B becomes:

network a ← → node S4 ← → S3 ← → S7 ← → S8 ← → network B.

Although the address refreshing scheme can solve the problem of MAC address errors caused by link switching, the message-based address refreshing scheme deletes all MAC addresses of nodes, so that instant network broadcast storms are caused, convergence time during switching is difficult to be less than 50ms due to the instant network broadcast storms, and network performance is greatly reduced. In fact, it is completely unnecessary to delete all the entries of the node address forwarding table, and the erroneous forwarding entries on the ring network are only those of the data flows intercepted by the new blocking point, and the forwarding entries of most of the data flows should not be deleted.

From the above analysis, it can be seen that the existing method of refreshing addresses using protocol messages can cause an instant broadcast storm of a multi-ring network. Therefore, it is very meaningful to provide a new type of selective address refresh scheme.

Disclosure of Invention

In view of the above, the present invention provides a method and a system for selectively refreshing an address of an ethernet multi-ring network, which can improve the performance of the ethernet multi-ring network.

In order to achieve the purpose, the technical scheme of the invention is realized as follows:

a method for selective address refresh for ethernet multiple ring networks, the method comprising: the Ethernet multi-ring network adopts the data message of the loop back to realize the selective address refreshing; the looped back data messages entering the ring are only propagated on the ring.

Wherein, when the ring node is a non-interconnected node, the method further comprises: the node on the ring detects that the adjacent link fails and blocks the port on the ring connected with the link; all address forwarding entries associated with the ring upper port are deleted simultaneously.

Wherein, when the ring node further comprises a blocked ring port, the method further comprises: and the ring node receives the data message, searches an address forwarding table by using a Virtual Local Area Network (VLAN) and a destination media access control address (DMAC) at the head of the data message, and loops the data message back if a corresponding address forwarding item is not searched.

Wherein, under the condition of the data message loopback, the method further comprises the following steps: packaging the outer layer of the looped data message by a looped VLAN, and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

Wherein, under the condition that the ring node receives the looped data message, the method also comprises: the ring node forwards the looped data message from another non-blocked ring port in a multicast mode; copying a copy of the looped data message for the analysis of the nodes on the loop;

wherein the analyzing comprises: and the ring node searches an output port of the looped data message in an address forwarding table by adopting the inner layer VLAN and the DMAC of the looped data message head.

Wherein, when the ring node is an interconnection node, the method further comprises: the node on the ring detects that the adjacent link fails and blocks the port on the ring connected with the link; and deleting the address forwarding entries associated with the ports on the ring.

Wherein, when the ring node further comprises a blocked multi-ring port, the method further comprises: and the ring node receives the data message, searches an address forwarding table by using the VLAN and the DMAC at the head of the data message, and loops the data message back if a corresponding address forwarding entry is not searched.

Wherein, under the condition of the data message loopback, the method further comprises the following steps: packaging the outer layer of the looped data message by a looped VLAN;

the loopback mode adopted by the loopback data message comprises the following steps: if the blocking port of the node on the ring is on the ring, the looped data message is forwarded from the port on the other ring; or, if the blocked port of the ring node is on the subring, the looped data message is forwarded from other non-blocked ring ports.

Wherein, under the condition that the ring node receives the looped data message, the method also comprises: the ring node forwards the looped data message from a non-blocked ring upper port; copying a copy of the looped data message for the analysis of the nodes on the loop;

A selective address refresh system for ethernet multiple ring networks, the system comprising: a refresh unit and a propagation unit; wherein,

the refreshing unit is used for realizing selective address refreshing by adopting a looped data message in the Ethernet multi-ring network;

and the transmission unit is used for transmitting the looped data message entering the ring only on the ring.

Wherein, this system still includes: a blocking and deleting unit, configured to detect that an adjacent link fails when a ring node is a non-interconnected node, and block a ring port connected to the link; all address forwarding entries associated with the ring upper port are deleted simultaneously.

Wherein, this system still includes: the packaging and forwarding unit is used for packaging the outer layer of the looped data message by a loopback VLAN under the condition of loopback of the data message and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

Wherein, this system still includes: a blocking and deleting unit, configured to detect that an adjacent link fails and block an on-ring port connected to the link when an on-ring node is an interconnection node; and deleting the address forwarding entries associated with the ports on the ring.

Wherein, this system still includes: and the loopback unit is used for receiving the data message and searching an address forwarding table by using the VLAN and the DMAC at the head of the data message when the ring node also contains a blocked multi-ring port, and looping the data message back if a corresponding address forwarding item is not searched.

Wherein the loopback unit further comprises: the device comprises a packaging module and a loopback mode selection module; wherein,

the packaging module is used for packaging the outer layer of the looped data message by a loopback VLAN under the condition of the loopback of the data message;

the loopback mode selection module is used for selecting to forward the looped data message from the port on the other ring when the blocking port of the ring node is on the ring; or, when the blocking port for the ring node is on the sub-ring, the looped data message is selected to be forwarded from the ports on other non-blocking rings.

Wherein, this system still includes: the forwarding and analyzing unit is used for forwarding the looped data message from the non-blocked ring upper port under the condition that the looped data message is received by the ring upper node; copying a copy of the looped data message for the analysis of the nodes on the loop;

The Ethernet multi-ring network adopts the looped data message to realize the selective address refreshing; the looped-back data messages entering the ring are only propagated on the ring.

By adopting the invention, under the scene of Ethernet multi-ring network, the address refreshing is selectively realized through the looped data message, thereby avoiding the instant broadcast storm caused by adopting protocol messages to refresh all addresses in the prior art, and further improving the performance of the Ethernet multi-ring network.

Drawings

FIG. 1 is a schematic diagram of an Ethernet multi-ring network;

FIG. 2 is a schematic diagram of Ethernet multi-ring network protection;

FIG. 3 is another schematic diagram of Ethernet multiple ring network protection;

FIG. 4 is yet another schematic diagram of Ethernet multiple ring network protection;

FIG. 5 is a diagram of a specific solution for address refresh by multiple rings in G.8032;

FIG. 6 is another diagram of a specific solution for address refresh by multiple rings for G.8032;

FIG. 7 is a flowchart of a first embodiment of the method of the present invention;

FIG. 8 is a flowchart of a second embodiment of the method of the present invention;

FIG. 9 is a schematic view of example one of the present invention;

FIG. 10 is another schematic illustration of a first embodiment of the present invention;

FIG. 11 is yet another schematic illustration of a first embodiment of the present invention;

FIG. 12 is a schematic view of example two of the present invention;

FIG. 13 is another schematic diagram of example two of the present invention.

Detailed Description

The basic idea of the invention is: the Ethernet multi-ring network adopts a looped data message to realize selective address refreshing; the looped-back data messages entering the ring are only propagated on the ring.

The following describes the embodiments in further detail with reference to the accompanying drawings.

A method for selective address refresh for ethernet multiple ring networks, the method comprising: the Ethernet multi-ring network adopts a looped data message to realize selective address refreshing; the looped-back data messages entering the ring are only propagated on the ring.

The following describes the processing cases of the ring nodes being non-interconnected nodes or interconnected nodes, respectively. Here, it should be noted that: when the ring node is an interconnect node, it is necessary to specify the direction from which the data packet is forwarded, that is, the interconnect node is required to control the forwarding direction of the data packet, unlike a non-interconnect node.

In the first case: the case where the nodes on the ring are non-interconnected nodes.

At this time, the method further includes: the ring node detects that the adjacent link fails and blocks a ring port connected with the link; all address forwarding entries associated with ports on the ring are deleted at the same time.

Here, when the ring node contains a blocked ring port in addition to the non-interconnected node, the method further comprises: and the ring node receives the data message, searches an address forwarding table by using the VLAN at the head of the data message and a destination media access control address (DMAC), and loops the data message back if a corresponding address forwarding entry is not found.

Here, in the case of loopback of the data packet, the method further includes: packaging the outer layer of the looped data message by a looped VLAN, and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

Here, in the case that the ring node receives the looped data packet, the method further includes: the ring node transmits the data message of the ring back from another non-blocked ring upper port in a multicast mode; and simultaneously copying a looped data message for the analysis of the nodes on the loop.

Wherein the analyzing comprises: and the ring upper node searches an output port of the looped data message in the address forwarding table by adopting an inner layer VLAN and a DMAC of the looped data message head.

In the second case: the ring nodes are interconnected nodes.

At this time, the method further includes: the ring node detects that the adjacent link fails and blocks a ring port connected with the link; while deleting the address forwarding entries associated with the ports on the ring.

Here, when the ring node contains a blocked multi-ring port in addition to being an interconnect node, the method further includes: and the ring node receives the data message, searches an address forwarding table by using the VLAN and the DMAC at the head of the data message, and loops the data message back if a corresponding address forwarding entry is not searched.

Here, in the case of loopback of the data packet, the method further includes: and packaging the outer layer of the looped data message by a looped VLAN.

One loopback mode adopted by the looped data message comprises: if the blocking port of the node on the ring is on the ring, the looped data message is forwarded from the port on the other ring; or, another loopback mode adopted by the looped data message comprises: if the blocked port of the node on the ring is on the subring, the looped data message is forwarded from other non-blocked ports on the ring.

Here, in the case that the ring node receives the looped data packet, the method further includes: the ring node forwards the looped data message from the non-blocked ring upper port; and simultaneously copying a looped data message for the analysis of the nodes on the loop.

In summary, the present invention mainly includes the following contents:

the core content of the invention mainly comprises: when the Ethernet multi-ring network fails or recovers, the ring network realizes fast switching by using the looped data message; the Ethernet multi-ring network carries out selective address refreshing by using the looped data message, thereby avoiding the broadcast storm of the whole Ethernet multi-ring network; the looped data packet will only propagate on the ring once it enters the ring (or main ring).

To achieve the core content described above, an ethernet multi-ring network node must contain the following functions, discussed in two cases, a normal ring-on node and an interconnect node:

firstly, the method comprises the following steps: when a node is a normal ring node (non-interconnected node), the node should contain the following functions:

firstly, the method comprises the following steps: when a node detects a failure of an adjacent link, the node blocks the ports on the ring adjacent to the link. All forwarding entries associated with that port are deleted at the same time.

II, secondly: when a data message received by a node (including a blocked port on the ring) searches a forwarding table by using the VLAN and the DMAC of the head part of the data message, if the corresponding entry is not searched, the data message is looped back.

Thirdly, the method comprises the following steps: the looped data packets must be encapsulated by the looped VLANs and forwarded on the ring in multicast. Wherein, the whole multi-ring network only has one loopback VLAN.

For the loopback VLAN, the loopback VLAN is used to identify that the current data packet is a looped data packet, that is, if the current data packet is encapsulated by the loopback VLAN, it is verified that the current data packet is a looped data packet. In other words, the looped-back data packet is a data packet encapsulated by the looped-back VLAN, and the looped-back data packet and the looped-back VLAN are all the meanings herein, which are not described in detail.

Fourthly, the method comprises the following steps: when the node receives the looped data message, the node forwards the data message from another non-blocked ring upper port in a multicast mode, and copies one part for the node to analyze. The node uses inner layer VLAN and DMAC of the looped data message to search a forwarding table, and processes the following conditions:

1. and if the found output port is the port on the ring for receiving the looped data message, the node deletes the found address forwarding entry.

2. If the searched output port is the lower ring port, the ring-up node strips the loopback VLAN of the looped data message, and then forwards the message from the corresponding lower ring port.

3. If the found output port is another ring port (not the ring port receiving the data message), the ring node does not perform relevant processing.

4. If the corresponding address forwarding entry is not found in the address forwarding table, the node does not perform any processing on the address forwarding table.

Secondly, the method comprises the following steps: when a node is an interconnect node, the node should contain the following functions:

firstly, the method comprises the following steps: when a node detects a failure of an adjacent link, the node blocks an on-ring port adjacent to the link while deleting a forwarding entry associated with the port.

II, secondly: when the interconnection node (including the multi-ring port with the block) receives the data message and looks up the address forwarding table with the VLAN and DMAC of the header of the data message, if the corresponding address forwarding entry is not found, the data message will be looped back. The looped data message must be encapsulated by the looped VLAN, and the loop-back mode is divided into the following two cases:

1. if the blocked port is on the ring (or main ring), the looped data message is forwarded from the port on the other ring.

2. If the blocked port is on the subring, the looped data message is forwarded from other non-blocked ports on the subring

Thirdly, the method comprises the following steps: when the interconnection node receives the looped data message, the node forwards the looped data message from the non-blocked ring (or main ring) port, and copies the data message to the interconnection node for analysis. The node uses the inner layer VLAN and the DMAC of the message to search an address forwarding table, and the following conditions are divided for discussion:

1. if the found port is the port for receiving the message, the interconnection node deletes the address forwarding entries corresponding to the inner layer VLAN and the DMAC of the message.

3. If the port found is a multi-ring port that is not receiving the packet (not blocked), the node does nothing to the address forwarding table.

The invention is illustrated below. Here, it should be noted that: the flow of the following method embodiment only illustrates the generation of the looped data message and how to delete the address by using the looped data message; the forwarding of messages to a normal ethernet follows the prior art.

The first embodiment of the method comprises the following steps: as shown in fig. 7, the data processing flow of the non-interconnected node of this embodiment mainly includes the following steps:

step 401, determine whether the node is an interconnect node. If not, proceed to step 402.

Step 402, judging whether the adjacent link of the node is failed. If yes, go to step 403; otherwise, go to step 404.

Step 403, the node blocks the data forwarding function of the multi-ring port adjacent to the failed link, and at the same time, deletes all address forwarding entries associated with the port, and then proceeds to step 404.

Step 404, the node receives the data message, and analyzes whether the VLAN is a loopback VLAN, that is: judging whether the received data message is a looped data message, if not, turning to the step 405; otherwise, go to step 411.

Step 405, the node uses the VLAN and DMAC of the data message header to find an egress port of the message in a forwarding table.

Step 406, determine whether a forwarding entry is found. If so, go to step 407; otherwise, go to step 408.

Step 407, the node forwards the data packet from the searched port.

Here, the found port may be only the lower ring port.

Step 408, the node broadcasts the data message to other ports except for the port receiving the data message.

Step 409, judging whether a node has a blocked port on the ring, if so, turning to step 410.

Step 410, the node loops back the data message from the port on the other ring, and the looped data message is encapsulated to loop back to the VLAN.

Step 411, after receiving the looped data message, the node forwards the looped data message from another non-blocked ring upper port; at the same time, one copy is copied for node analysis.

Here, during analysis, the node looks up the forwarding table using the inner layer VLAN and DMAC of the looped data packet.

And step 412, processing in 4 cases.

Here, case 1: the found port is the port receiving the message, and the process proceeds to step 413.

Case 2: the found egress port is the lower ring port and the process proceeds to step 414.

Case 3: the found port is a multi-ring port that is not receiving the message (not blocked) and proceeds to step 415.

Case 4: if no corresponding forwarding entry is found, proceed to step 416.

Step 413, the node deletes the found forwarding entry.

Step 414, the node strips the loopback VLAN of the looped data packet, and then forwards the packet from the corresponding lower ring port.

Step 415, the node does not process the forwarding table.

Step 416, the node does not process the forwarding table.

The second method embodiment: as shown in fig. 8, the data processing flow of the interconnection node of this embodiment mainly includes the following steps:

step 501, judging whether the node is an interconnection node. If so, proceed to step 502.

Step 502, determining whether the adjacent link of the interconnection node fails. If yes, go to step 503; otherwise, go to step 504.

Step 503, the interconnect node blocks the data forwarding function of the multi-ring port adjacent to the failed link, and at the same time, deletes all address forwarding entries associated with the port, and then proceeds to step 504.

Step 504, the interconnection node receives the data message, and analyzes whether the VLAN is a loopback VLAN, that is: judging whether the received data message is a looped data message, if not, turning to step 505; otherwise, go to step 511.

Step 505, the interconnection node uses the VLAN and DMAC of the data packet header to find an egress port of the packet in the forwarding table.

Step 506, determine whether a forwarding entry is found. If so, go to step 507; otherwise, go to step 508.

Step 507, the interconnection node forwards the data message from the searched port.

At step 508, the interconnect node broadcasts the data message to ports other than the port receiving the message.

Step 509, determine whether there is a blocked on-ring port on the interconnect node, if so, go to step 510.

Step 510, the interconnection node loops the data packet to an upper ring port that is not blocked on the ring (or called the primary ring), and the looped data packet is encapsulated to loop back to the VLAN.

Step 511, after receiving the looped data message, the interconnect node forwards the looped data message from the ring upper port of the non-blocking ring (or called main ring); and a copy is made to the interconnect node for analysis.

Here, during the analysis, the interconnection node searches a forwarding table by using the inner layer VLAN and DMAC of the packet.

And step 512, processing in 4 cases.

Here, case 1: the found port is the port receiving the message, and the process proceeds to step 513.

Case 2: the found egress port is the lower ring port and proceeds to step 514.

Case 3: the found port is a multi-ring port that is not receiving the message (not blocked) and proceeds to step 515.

Case 4: if no corresponding forwarding entry is found, proceed to step 516.

Step 513, the node deletes the found forwarding entry.

Step 514, the node strips the loopback VLAN of the looped data packet, and then forwards the packet from the corresponding lower ring port.

Step 515, the node does not process the forwarding table.

Step 516, the node does not process the forwarding table.

The following further illustrates the selective address refresh scheme based on the loopback data message by examples one and two, but not limiting the invention.

Example one: link switching in case of failure of a ring (or called main ring).

As shown in fig. 9, Ring1 is a complete closed Ring, and includes nodes such as S1, S2, S3, S4, S5, and S6, and links therebetween, where the node S2 is a control node, and it is a forwarding function that blocks data packets of 21 ports and Ring2 protocol packets (the Ring2 protocol packet is treated as a data packet in the Ring 1) under the condition that Ring1 has no fault; ring2 is a sub-Ring, which contains S1, S8, S7 and S3 nodes, S7 node is a control node, which blocks the data packet forwarding function of 72 port under the condition that Ring2 has no fault, but does not block the protocol packet on the control VLAN of Ring 2. In the case of no failure of the ethernet multi-ring network, the address entries in the address forwarding tables of the nodes on the ethernet multi-ring network reaching network a and network B are as shown in fig. 9, and the communication paths between network a and network B are:

As shown in fig. 10, when the link < S6, S5> of Ring1 fails, and the nodes S6 and S5 detect the failure, the nodes S6 and S5 block the data forwarding functions of the

ports

61 and 52 and delete the address forwarding entries associated with them, and simultaneously send SF protocol packets periodically along the two ports. After receiving the SF message, the control node S2 opens the data message forwarding function of the 21 port. The refresh scheme of the present invention will be described first by taking the data flow of network B → network a as an example.

When the network B sends the data message P to the network A, after the data message P reaches the node S6 along the path before the fault occurs, the node S6 (with the upper fault port) searches an address forwarding table by using the VLAN1 and the MAC (A) at the head of the data message P, and does not find a corresponding forwarding entry, and the node S6 loops the data message and simultaneously packages the upper loop VLAN; after receiving the looped data packet P, the node S1 (interconnection point) first forwards the packet P from another main ring port 12, and then copies one packet for analysis by the S1 node. The node S1 uses the inner layer VLAN and MAC (A) of the looped data message P to search the forwarding table, finds that the searched port is consistent with the port of the received looped data message P, and deletes the forwarding items corresponding to the VLAN1 and the MAC (A); after receiving the looped-back data packet P, the nodes S2, S3, S4 and S5 forward the packet from another non-blocked main ring port, and copy it for analysis. The nodes use the inner layer VLAN and MAC (A) of the looped data message P to search the forwarding table, and find that the searched port is the other primary ring port of the node, and the nodes do not process the forwarding table.

In addition, when the network a sends the data packet M to the network B, the method for processing the packet M and deleting the address by using the loopback packet M (encapsulating the special VLAN) is the same as the method for processing the data packet P by the system, and the specific process is shown in fig. 11.

After a period of time, the nodes on the multiple rings learn the MAC address of network a on VLAN1 and network B on VLAN 2. The communication path between network a and network B becomes:

network a ← → node S4 ← → S3 ← → S2 ← → S1 ← → S8 ← → network B.

Example two: and switching the link when the subring fails.

As shown in fig. 12, when the link < S1, S8> of Ring1 fails, and the nodes S1 and S8 detect the failure, the nodes S1 and S8 block the data forwarding functions of the

ports

13 and 81 and delete the address forwarding entries associated with them, respectively, while periodically sending SF protocol packets along the two ports, respectively. After receiving the SF message, the control node S7 opens the data message forwarding function of the port 72. The method of the present invention will be described below by taking the data flow of network B → network a as an example.

When the node S8 (with the ring fault port) receives the data packet R sent from the network B (sent to the network a on the VLAN 1), the VLAN1 and the mac (a) at the header of the packet R are used to look up a table, and it is found that there is no corresponding entry, and the node S8 loops back the data packet R and encapsulates the looped VLAN.

After receiving the looped data message R, the node S7 first forwards the message R from another sub-ring port 72, and then copies a copy for analysis by the S7 node. The node S7 checks the forwarding table with the inner VLAN and mac (a) of the looped data packet R, finds that the checked port is consistent with the port of the received looped data packet R, and deletes the forwarding entry corresponding to the VLAN1 and mac (a).

After receiving the looped data message R, the node S3 (interconnection point) first forwards the message R from the two

main ring ports

32 and 31, and then copies one for analysis by the S3 node. The node S3 checks the forwarding table with the inner VLAN and mac (a) of the looped data packet R, finds that the checked port is inconsistent with the port of the received looped data packet R, and does not process the forwarding entries corresponding to the VLAN1 and mac (a).

After receiving the looped-back data message R from the nodes S5, S6, S1, and S2, the node forwards the message from another non-blocked main ring port, and copies the message for analysis. The nodes use the inner layer VLAN1 and MAC (A) of the looped data message R to search the forwarding table, find that the searched port is consistent with the port receiving the message, and delete the corresponding entries of the VLAN1 and the MAC (A) in the forwarding table.

After receiving the looped data message R, the node S4 forwards the message from the port 42 on the other primary ring, and copies it for analysis. The node S4 uses the inner VLAN1 and mac (a) of the looped data packet R to search the forwarding table, finds that the searched port 43 is a lower ring port, and does not process the forwarding entries corresponding to the VLAN1 and mac (a). And simultaneously, the outer-layer loopback VLAN of the loopback data message R is stripped, and the processed message is forwarded out from the 43 port.

In addition, when the network a sends the data packet Q to the network B, the method for processing the packet Q and deleting the address by using the loopback packet Q (encapsulating the special VLAN) is the same as the method for processing the data packet R by the system, and the specific process is shown in fig. 13.

network a ← → node S4 ← → S3 ← → S7 ← → S8 ← → network B.

Through the above analysis, the present invention well solves the difficulties encountered when the existing mechanism of address refreshing by protocol message is applied on the ethernet multi-ring network, and greatly improves the performance of the ethernet multi-ring network.

A selective address refresh system for ethernet multiple ring networks, the system comprising: a refresh unit and a propagation unit. The refreshing unit is used for realizing selective address refreshing by adopting a looped data message in the Ethernet multi-ring network. And the transmission unit is used for transmitting the looped data message entering the ring only on the ring.

The specific implementation of each unit in the system is different when the ring node is a non-interconnected node or an interconnected node, which is described below.

The system further comprises: a blocking and deleting unit, configured to detect that an adjacent link fails and block an on-ring port connected to the link when the on-ring node is a non-interconnected node; all address forwarding entries associated with ports on the ring are deleted at the same time.

Here, the system further includes: the packaging and forwarding unit is used for packaging the outer layer of the looped data message by a loopback VLAN under the condition of loopback of the data message and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

In the second case: the ring nodes are interconnected nodes.

The system further comprises: a blocking and deleting unit, configured to detect that an adjacent link fails and block an on-ring port connected to the link when the on-ring node is an interconnection node; while deleting the address forwarding entries associated with the ports on the ring.

Here, the system further includes: and the loopback unit is used for receiving the data message and searching the address forwarding table by using the VLAN and the DMAC at the head of the data message when the ring node also contains a blocked multi-ring port, and looping the data message if the corresponding address forwarding item is not searched.

Here, the loopback unit further includes: the device comprises an encapsulation module and a loopback mode selection module. The packaging module is used for packaging the outer layer of the data message in a loopback VLAN under the condition of loopback of the data message. A loopback mode selection module, which is used for selecting to forward the looped data message from the port on the other ring when the blocking port of the ring node is on the ring; or, when the blocking port for the ring node is on the subring, the looped data message is selected to be forwarded from other non-blocking ring ports.

Here, the system further includes: the forwarding and analyzing unit is used for forwarding the looped data message from the non-blocked ring upper port under the condition that the looped data message is received by the ring upper node; and simultaneously copying a looped data message for the analysis of the nodes on the loop.

Here, the above chinese and english related characters are explained by reference: the Ring is represented by Ring; the subloop is represented by Sub-Ring; the Interconnection Node is represented by an Interconnection Node; the address forwarding table is also called FDB table; loopback is represented by wrap; DMAC refers to the destination MAC address, where MAC refers to media access control;

the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.

Claims

1. A selective address refresh method for ethernet multiple ring networks, the method comprising: the Ethernet multi-ring network adopts a looped data message to realize selective address refreshing; the looped data message entering the ring is only transmitted on the ring;

when the ring node is a non-interconnected node, the method further comprises: the node on the ring detects that the adjacent link fails and blocks the port on the ring connected with the link; simultaneously deleting all address forwarding entries associated with the ports on the ring;

when the ring node further includes a blocked ring port, the method further includes: the ring node receives a data message, searches an address forwarding table by using a Virtual Local Area Network (VLAN) and a destination media access control address (DMAC) at the head of the data message, and loops back the data message if a corresponding address forwarding item is not searched;

under the condition of the data message loopback, the method further comprises the following steps: packaging the outer layer of the looped data message by a looped VLAN, and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

2. The method of claim 1, wherein in the event that the looped-back data packet is received by the node on the loop, the method further comprises: the ring node forwards the looped data message from another non-blocked ring port in a multicast mode; copying a copy of the looped data message for the analysis of the nodes on the loop;

3. The method of claim 1, wherein when the ring node is an interconnect node, the method further comprises: the node on the ring detects that the adjacent link fails and blocks the port on the ring connected with the link; and deleting the address forwarding entries associated with the ports on the ring.

4. The method of claim 3, wherein when the on-ring node further contains a blocked multi-ring port, the method further comprises: and the ring node receives the data message, searches an address forwarding table by using the VLAN and the DMAC at the head of the data message, and loops the data message back if a corresponding address forwarding entry is not searched.

5. The method of claim 4, wherein in case of loopback of the data packet, the method further comprises: packaging the outer layer of the looped data message by a looped VLAN;

6. The method of claim 5, wherein in the event that the looped-back data packet is received by the node on the loop, the method further comprises: the ring node forwards the looped data message from a non-blocked ring upper port; copying a copy of the looped data message for the analysis of the nodes on the loop;

7. A selective address refresh system for ethernet multiple ring networks, the system comprising: a refresh unit and a propagation unit; wherein,

a transmission unit, configured to transmit the looped data packet entering the ring only on the ring;

the system further comprises: a blocking and deleting unit, configured to detect that an adjacent link fails when a ring node is a non-interconnected node, and block a ring port connected to the link; simultaneously deleting all address forwarding entries associated with the ports on the ring;

the system further comprises: a loopback unit, configured to, when the ring node further includes a blocked multi-ring port, receive a data packet and search an address forwarding table by using a VLAN and a DMAC at a header of the data packet, and if a corresponding address forwarding entry is not found, loop back the data packet;

the system further comprises: the packaging and forwarding unit is used for packaging the outer layer of the looped data message by a loopback VLAN under the condition of loopback of the data message and forwarding the looped data message on a ring in a multicast mode; wherein, the whole Ethernet multi-ring network only has one loopback VLAN.

8. The system of claim 7, wherein the loopback unit further comprises: the device comprises a packaging module and a loopback mode selection module; wherein,

the loopback mode selection module is used for selecting to forward the looped data message from the port on the other ring when the blocking port of the ring node is on the ring; or, when the blocking port for the ring node is on the subring, the looped data message is selected to be forwarded from the ports on other non-blocking rings.

9. The system of claim 8, further comprising: the forwarding and analyzing unit is used for forwarding the looped data message from the non-blocked ring upper port under the condition that the looped data message is received by the ring upper node; copying a copy of the looped data message for the analysis of the nodes on the loop;