JP2003258822A - Packet ring network and inter-packet ring network connection method used in the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a packet ring network capable of connecting a plurality of rings and performing high-speed switching even when the plurality of nodes connect the rings for the formation of redundancy. <P>SOLUTION: An active node determining means 121 of a linking node for making connection between rings determines whether or not a self-node should be a node for transferring a packet between rings by mutually confirming the states of both nodes. A node (active node) for transferring the packet passes the packet between rings 101 and 102 through ring packet blocking means 123 and 124, and a node (standby node) that does not transfer the packet does not pass the packet between the rings 101 and 102 with the ring packet blocking means 123 and 124. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a packet ring network and a connection method between packet ring networks used therein, and more particularly to a connection method between rings in a packet ring network.

[0002]

2. Description of the Related Art Conventionally, when configuring a network using Ethernet (R), a spanning tree protocol has been used for redundancy. Spanning tree is a topology with the feature that there are no loops in the network topology.

Regarding the protocol for constructing the above-mentioned spanning tree, IEEE802.1D MAC bridges is used as the spanning tree protocol.
Has been standardized as a specification. When this protocol is used, when a bridge in operation fails, the port of the bridge that was blocked to eliminate the loop starts to operate, which is advantageous in that a redundant configuration can be built.

It has been pointed out that sending a broadcast to a topology with loops causes a so-called broadcast storm, where the broadcast packets are duplicated on the network. However,
If it becomes a tree, any LA in the entire topology
It can be guaranteed that there is only one copy of the packet recognized on N (Local Area Network). This is because there is no loop on the network, and there is only one route to a certain destination.

In the spanning tree protocol, when a bridge failure or the like occurs, all the trees are reconfigured, so that data transmission / reception is not performed until the tree creation is completed, and there is a problem that it takes time for the network to converge. This may take a few minutes or more, depending on the network size and the number of connected bridges.

In addition, RPR (Resilient Pa
When a spanning tree is applied to a ring network such as a Ccket Ring), there is a fundamental problem that the ring that is a communication path is blocked and the intended band cannot be effectively used.

Further, a routing protocol implemented in a router is used to make the network redundant at Layer 3. When the topology changes,
The router detects the change and recalculates the routing table so that it will be reflected in the new topology.

IP (Internet Protocol)
1) RIP (Routing Information), which is the most commonly used routing protocol.
Ion Protocol) and OSPF (Open S)
When using "hortest PathFirst", the convergence time from the change of the topology to the update becomes a problem.

A common problem of these techniques is that it takes several minutes or more for the protocol for redundancy to converge, and the correct transfer of packets is temporarily suspended until then.

Further, as a related prior art, for example, SRP (Spatial Reuse) manufactured by Cisco
Protocol). For SRP,
IETF (Internet Engineering)
RFC (Request of Task Force)
For Components) 2892's "The Ci
sco SRP MAC Layer Protoco
The technical contents are described in "1".

SRP is a technique for ring-connecting routers. Since the nodes are basically connected by layer 3, the rings are also connected by layer 3. Therefore,
RPR MAC (Media Access Cont)
The redundancy in the (roll) layer is not considered, and the route is switched by the routing protocol of the router.

[0012]

In the above-mentioned conventional packet ring network, in a ring network such as RPR, the duplication of a broadcast packet easily occurs unless a plurality of rings are connected by an appropriate method. It falls into a condition called broadcast storm.

Further, in the conventional packet ring network, redundant installation of network equipment is an essential condition in order to improve the reliability of the network. Therefore, measures against these broadcast packets and redundant installation are satisfied at the same time, and switching thereof is performed. Must be done correctly.

Therefore, an object of the present invention is to solve the above problems, enable a plurality of ring connections, and perform high-speed switching even when a plurality of nodes connect the rings for redundancy. It is to provide a packet ring network capable of performing the above and a connection method between the packet ring networks used for the packet ring network.

[0015]

A packet ring network according to the present invention comprises first and second contact nodes for connecting between first and second packet rings, and the first and second contact nodes are connected to each other. It is redundant.

A connection method between packet ring networks according to the present invention provides a connection between a first packet ring and a second packet ring.
Connections are made at the redundant first and second contact nodes.

That is, the connection method between packet ring networks of the present invention is highly expandable and reliable by connecting the packet rings by a plurality of nodes and making these contact nodes connecting the rings redundant. It is possible to provide a network.

In the present invention, of the nodes connecting the rings, usually only one node passes the packet between the rings so that the relay of the broadcast packet becomes one. Therefore, it is possible to prevent malfunction due to double reception of the same packet.

In addition, according to the present invention, when two nodes connecting between the rings are allowed to be active (a state in which packet passage is permitted) depending on the state of the fault, in the case of a fault in the ring. It also has the feature that it is possible to realize a flexible inter-ring connection that can be operated.

Further, according to the present invention, compared with the route change by the spanning tree protocol, the routing protocol, etc., the route switching is quick and the service interruption time can be shortened.

Further, according to the present invention, the MAC (Media) of the node that has become active from the forwarding table for each node in the ring that has become active.
By having the Access Control address deleted, the learning bridge is relearned quickly.

[0022]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the structure of a packet ring network according to an embodiment of the present invention. Referring to FIG. 1, a packet ring network according to an embodiment of the present invention includes a ring 10 for forwarding packets.
1, 102 is a contact node 12a connecting between the rings,
12b are connected to each other.

FIG. 2 is a block diagram showing the configuration of the contact nodes 12a and 12b of FIG. In FIG. 2, contact node 1
Since the configurations of 2a and 12b are the same, they are shown as the contact node 12.

The contact node 12 has an active node determination means 121 and an active / standby switching means 122.
And ring packet block means 123 and 124. In FIG. 2, the illustration of other functions of the contact node 12 as a node is omitted.

Connection nodes 12a, 12 for connecting the rings
The active node determination means 121 of b determines by mutual confirmation of the states of both nodes whether or not the own node should be a node that performs packet transfer between rings.

A node (active node) that transfers packets passes packets between the rings 101 and 102 through the ring packet block means 123 and 124,
The node (standby node) that does not perform the transfer is the ring packet blocking means 123, 124.
Since packets between No. 02 are not passed, it is possible to prevent a broadcast storm due to duplication of broadcast packets even in a topology having a loop as shown in FIG.

Depending on the state of the ring, both contact nodes 12a and 12b are activated to prevent double reception of a broadcast packet and to prevent double reception of both rings 101 and 10 even if a failure occurs in the ring.
Packet transfer between the two can be performed.

The active / standby switching means 122 of the contact nodes 12a and 12b is the active node determining means 121 based on whether or not it is possible to send and receive survival confirmation messages to each other using the rings 101 and 102 and failure information of the ring topology. The ring packet block means 123 and 124 are controlled so as to switch between active and standby according to the result of the determination.

Even when the active / standby switching of the contact nodes 12a, 12b is performed, the MAC address of the node which has been operating as active is registered in the transfer table in each node 11a, 11b in the ring.

In the conventional bridge or switch, the transfer process is performed according to the contents of the transfer table, and therefore the transfer is not performed correctly until the registration is deleted from the transfer table due to aging.

On the other hand, in this embodiment, when the switching operation is performed, each node 11 on the rings 101 and 102 is
A newly activated secondary node transmits a registration deletion message packet instructing the deletion of the transfer table registration addressed to the primary node to a and 11b, thereby deleting the unnecessary registration of the transfer table.

Next, the operation of this embodiment will be described.
This embodiment belongs to a ring network that transmits data packets on the ring. An example of such a network is currently IEEE 802.17 Working.
RPR (Resili) discussed in Group
ent Packet Ring). RP
R is a MAC layer protocol for providing a packet ring network, and is a technology under standardization for the purpose of effective use of network bandwidth, ensuring network reliability, and automatic operation by detecting ring topology. is there.

FIG. 3 is a block diagram showing a network configuration example of RPR, and FIG. 4 is a diagram showing a ring packet format used in the network configuration example shown in FIG. The network configuration of RPR and the operation of the node will be described with reference to FIGS. 3 and 4.

RPR is a ring network that uses a double ring topology that rotates in opposite directions in both directions. FIG. 3 shows an example of a network configuration in which four nodes 21a, 21b, 21c, 21d are connected to a ring. The double ring is generally called an inner ring 201 and an outer ring 202 as shown in FIG. 3, and both rings are used to transmit both data packets and control packets.

The control packet is used for network topology detection, bandwidth control, transmission direction switching at the time of failure, etc., but details thereof are not described here. Also, control packets are typically sent in the opposite direction to their corresponding data packets. That is, the control packet corresponding to the data packet transmitted by the outer ring 202 is the inner ring 20.
Sent at 1.

The format of the ring packet used is shown in FIG. Ring packets are basically Ethernet
It is similar to the layer 2 packet format such as et (R), but in order to transmit the packet in the ring, the destination node MAC address RDA, the source node MAC
Address RSA, identifies the type of packet (for example,
Ty for identifying data packets and control packets, etc.)
pe has a field called TTL (Time to live) for preventing packets from infinitely circulating in the ring.

In the format of the ring packet shown in FIG. 4, other fields such as a priority field indicating the priority of the packet are omitted. Further, the ring packet is transmitted on the ring by encapsulating user data such as Ethernet (R) from the user terminal.

Each node 21a, 21b, 21c, 21d
Performs the following packet receiving operation. In other words, the unicast packet is a ring packet destination node and is deleted from the ring when received.

The broadcast packet is received by each node and transferred to the next node. The packet goes around the ring and is removed from the ring when the broadcast packet returns to the source node. This prevents broadcast packets from going around the ring indefinitely.

Further, the TTL is decremented by 1 each time it is transferred to each node in the ring, and is deleted from the ring when the TTL becomes 0. This will cause the destination node to be removed from the ring, etc.
Prevents ring packets from reaching the destination and continuing infinitely around the ring.

Furthermore, the control packet is a packet used for ring protection, band control, topology detection, etc., but is received if necessary at each node. Since these details are not related to the present invention, description thereof will be omitted.

Each node 21a, 21b, 21c, 21d
Monitors the link state of the ring network connected to its own node, and when a failure occurs in the link, the node that detects the failure notifies the failure information in the ring by a ring packet. Based on this and the topology information of the ring, each node 21a, 21b, 21c, 21d
Can detect points of failure in the ring.

In the network as described above, in order to transmit the data from the user terminal 22a connected to the node 21c to a certain user terminal 22b, the destination user terminal 22b has to select which node (in FIG. 3) in the ring. Whether it is connected to the node 21b) must be registered in each node. This registration is not a manual setting,
To do this automatically, a method called a learning bridge is usually used.

The learning bridge is, for example, Ethernet.
In a layer 2 bridge network such as (R),
It is used to learn the packet receiving port and its source address in association with each other and to register them in the forwarding table.

In the ring network of this embodiment, the source address RSA of the received ring packet
And a method of learning by associating with the source address SA of the user data transmitting terminal will be referred to as a learning bridge. Further, even if the correspondence of the addresses is not learned by learning and the registration is performed by setting, no problem occurs in the operation.

In the network configuration shown in FIG. 3, when the learning bridge operates, the operation in each node is performed as follows. That is, the destination table is determined by referring to the transfer table using the destination address DA of the user data contained in the ring packet shown in FIG. 4 as the search key. At this time, the user data is encapsulated in a ring packet, the destination address RDA is the MAC address of the determined destination node, and the source address RSA is the MAC address of the own node.

As a result of searching the transfer table, if the destination node is unknown due to unlearning (if not registered)
Do flooding to. This is a normal Ethernet
It is almost the same as the flooding in et (R), etc., and is not flooded to the receiving port for user data. At this time, the destination address RDA is the broadcast address, and the source address RSA is the MAC address of the own node.

In this case, the broadcast packet is transmitted only in one direction of the ring. To determine the transmission direction of the ring, for example, the MA of the user terminal that is the transmission source is used.
It is possible to follow the result of passing the C address through an appropriate hash function, but this method will not be described in detail.

Further, as a result of searching the transfer table, when the MAC address RDA of the destination node can be determined, it is encapsulated in a ring packet and transmitted to the ring. At this time, the ring direction to be transmitted is usually the direction that is the shortest route to the destination node determined by another means (topology detection).

Furthermore, the node that receives the ring packet learns the MAC address of the source node of the ring packet, RSA, and the MAC address of the user terminal, SA, in association with each other, and registers them in the transfer table.

By the above operation, the MAC of the user terminal
It becomes possible for each node to learn the correspondence between the address and the node MAC address. Here, consider the case where two rings are connected by using two nodes. Hereinafter, the node that connects the rings is called a contact node. Since the contact node is used to connect a plurality of rings, it is considered that the node MAC address is assigned to each connected ring, but the operation is not hindered even if they are the same.

When the destination address RDA of the ring packet is a broadcast, two contact nodes flood the broadcast packet, so that the same packet is transmitted twice on the ring. Further, it is easily expected that a phenomenon called a broadcast storm will occur in which broadcast packets are not deleted forever.

The operation at this time will be briefly described with reference to FIG. The ring packet transmitted by the broadcast address from the node 11a goes around the ring 101 and is received by the contact node 12a. Since the destination address RDA of the received ring packet is broadcast, the contact node 12a receives the packet from the ring and transfers it to the next node.

Since the contact node 12a also transfers the packet to the other ring (the ring 102 in FIG. 1) to which it is connected, this broadcast packet is transmitted to the ring 1
It is also transferred to 02, and goes around in the ring 102. Furthermore, the ring packet of the broadcast address that circulates in the ring 101 is deleted from the ring 101 when it returns to the node 11a that is the transmission source.

At the time of transfer across rings by the contact node 12a, the address of the ring packet is once terminated,
The source address RSA of the broadcast packet transferred to the ring 102 is the node address of the contact node 12a. Therefore, this broadcast packet goes around the ring 102, is received by each node on the ring 102, and is deleted from the ring 102 when it returns to the contact node 12a.

Here, when the contact node 12b receives the broadcast packet, the node 12b is the contact node, so that the packet transfer process is also performed on the ring 101. Also at this time, the address of the ring packet is once terminated, so that the source address RSA is set in the ring 102.
, The ring packet of the contact node 12b goes around. This packet is transferred again to the ring 101 at the contact node 12a. In this way, a phenomenon called broadcast storm occurs in which broadcast packets are not deleted forever. This is the same even when the packet having the broadcast address is first received by the contact node 12b.

Similarly, the ring packet transmitted from the node 11a with the broadcast address is the ring 10
Since it circulates around 2 and is received by the contact node 12b, ring packets circulate as in the case of the contact node 12a, the broadcast packet is not deleted from the ring correctly, and the bandwidth is wasted. I will end up. Therefore, as it is, connect the rings with multiple contact nodes,
Network redundancy cannot be increased.

In order to avoid the above-mentioned malfunction, in this embodiment, the contact nodes 12a, 1 having a redundant configuration are used.
A loop of the broadcast packet is prevented by making it possible to transfer the packet between the rings only to the node that has transitioned to the active state in 2b. Here, in the present embodiment, transmission / reception is not blocked, and a node state in which packets can pass between the rings is defined as active, and a node state in which data packets are blocked so that they cannot pass is defined as standby.

FIG. 5 is a flow chart showing the operation of the connection nodes 12a and 12b of FIG. 1 when the device is activated. This Figure 5
The setting confirmation between the contact nodes at the time of starting the apparatus will be described with reference to. In this case, it is assumed that the own node is a contact node that has been notified by the setting, and the MAC address of the contact node that makes a pair is also given by the setting.

The setting of the contact node includes a setting as a primary node and a setting as a secondary node. Here, the primary node is a node set to be activated, and the secondary node is a node set to be made standby.

After the node is activated, both the primary node and the secondary node, which have been set as the contact nodes, confirm the contact node settings with each other. This confirmation is performed using a ring packet, and the destination address RDA of the ring packet
Are paired contact nodes, and the source address RSA is the MAC address of its own node. This confirmation is done to increase certainty and can be omitted in some cases.

After the node is activated, each node having the primary and secondary settings mutually transmits a ring packet describing the setting information of its own node to confirm the setting (step S1 in FIG. 5). When the mutual settings are confirmed between adjacent nodes and there is no setting error (step S2 in FIG. 5),
The primary node releases (activates) the block of transmission and reception to both rings (steps S3 and S in FIG. 5).
4). The secondary unblocks only the transmission / reception of one side ring of both connected rings (standby) (steps S3 and S5 in FIG. 5).

However, the ring packet to be blocked is only the data packet, and the control packet and the survival confirmation packet for detecting the failure between the contact nodes are not blocked. This can be easily identified by the Type field in the ring packet format shown in FIG. At this time, the ring for releasing the block is determined, for example, by the size of the ring ID assigned in advance.
In addition, the preset ring side may be released.

In order for the secondary node to transit from the standby state to the active state, (1) when a failure of the active primary node is detected, (2)
When it is possible to determine from the failure information sent by each node when there is a failure in the ring that there are no multiple routes to reach a certain node, for example, if the existence of the primary node cannot be confirmed via one ring, It depends on the condition of a failure state (including a case where it is determined that the communication path with the primary node is disconnected).

FIG. 6 is a diagram showing an example of transition of the secondary node from the standby state to the active state in the ring network shown in FIG. FIG. 6 shows an example in the case of the above condition (2).

In this case, the failure points 111 and 112 prevent the contact confirmation messages from being exchanged between the contact nodes 12a and 12b via the ring 102. However,
Since it is performed via the ring 101, the contact node 12b that is on standby does not become active under the condition (1) alone.

Therefore, if the operation is performed as it is, the ring packet from the node 11a is transferred to the ring 102 only through the active contact node 12a. This ring packet is the failure point 1 of the ring 102.
11 and 112 cannot be received by the node 11d.
Therefore, it can be seen that the necessity of activation arises even when the existence confirmation between the adjacent nodes is performed via the ring 101. That is, the secondary node may be activated at the same time under the condition other than the detection of the failure of the primary node.

FIG. 7 is a flow chart showing a process of activating a contact node when another node fails according to an embodiment of the present invention. With reference to FIG. 7, a process of activating a contact node when another node fails according to an embodiment of the present invention will be described.

The primary node and the secondary node mutually transmit survival confirmation packets. For transmitting and receiving this survival confirmation packet, the inner ring and the outer ring of both rings to which both nodes are connected are periodically transmitted (step S11 in FIG. 7).

Further, in this survival confirmation packet, the ID of the ring to be transmitted and the direction of the ring to be transmitted (inner ring or outer ring) are specified, and which packet is the response to the survival response packet. By clearly indicating, it is possible to easily identify whether it is a response to the inner ring or a response to the outer ring.

After transmitting the survival confirmation packet, each node waits for reception of a survival response packet from a contact node that is a pair, which is a response to the survival confirmation packet (step S12 in FIG. 7). When the alive response packet is not received within the preset timeout time T1, the number of times of the reception time timeout is checked (step S13 in FIG. 7).
In this check, if the number of timeouts is less than the predetermined number of retransmissions N1, the state transits to the state of transmitting the survival confirmation packet again (return to step S11). This confirmation of the number of retransmissions is performed for each transmission direction of the ring ID and the survival confirmation packet. Therefore, each node holds a state for each ring ID and transmission direction of the survival confirmation packet.

When the survival response packet is received within the timeout time T1, each node confirms whether a failure has occurred in the ring (step S14 in FIG. 7). this is,
As described above, the operation of each node has already been notified from another node. If a failure has occurred in the ring, it is determined whether both nodes can be activated (step S15 in FIG. 7). When it can be determined that the own node can be activated, the state of the own node is transitioned to activated (step S16 in FIG. 7).

If there is no fault in the ring in step S14, or if it is determined in step S15 that the ring cannot be activated, this flow is ended, and the transmission timing of the next survival confirmation packet is awaited.

If the number of times of timeout exceeds the predetermined number of times of retransmission N1 in step S13, the reception state of the survival confirmation packet in the own node is confirmed (step S17 in FIG. 7). As a result of the confirmation, if the timeout of the survival confirmation packet via all the routes is equal to or greater than the number of retransmissions, it is determined that a failure has occurred in the paired contact node, and the process proceeds to step S18. Step S
In 17, the number of times of timeout is the number of retransmissions N
If there is a route that is less than 1, it is not determined that a failure has occurred in the contact node, and the process returns to step S11.

When it is determined that a failure has occurred in the contact node, the state of the own node is confirmed (step S1 in FIG. 7).
8). If the status of the local node is standby,
Judging that the active node has failed,
The own node is activated (step S19 in FIG. 7) and the block is released. If the state of the own node is active, the active state is maintained (step S20 in FIG. 7).

Even after the active / standby switching of the contact node is performed, the MAC address of the primary node is registered in the transfer table as a result of learning in each node in the ring. Therefore, this registration must be deleted by some means in order for the packet between the rings to pass through the secondary node.

In the conventional bridge or switch, the above registration is deleted from the transfer table due to the passage of time called aging. Therefore, until the registration of the transfer table is deleted by aging, there is an adverse effect that it is transmitted to the wrong destination.

Therefore, in this embodiment, when the contact node is switched, a registration deletion message packet for instructing the registration deletion addressed to the contact node paired with the active secondary node is sent to each node in the ring. Sending facilitates deregistration from the forwarding table. When transmitting to each node in the ring, the broadcast address is used for effective use of the band, but other means may be used.

The node receiving the registration deletion message packet deletes the entry addressed to the registration deletion message from the transfer table of its own node according to the message content. As a result, the packet addressed to the failed node is relearned, and this time, the inter-ring connection is performed via the secondary node.

In the present embodiment, an example of connecting two rings is shown, but according to the same connection method, it is also applicable to connecting three or more rings. Further, according to the present embodiment, by connecting the rings with a plurality of nodes, it is possible to construct a highly reliable network as compared with the case of connecting with one node. When a failure occurs in one node, the spare node operates
There is an advantage that a network that is resistant to failures can be configured.

In this embodiment, since the survival confirmation packet is transmitted and received only between the contact nodes that connect the rings, a spanning tree protocol that requires rebuilding the spanning tree of the entire network and a complicated link state algorithm are used. Compared with the routing protocol used, the convergence time due to network failure can be greatly shortened when a redundant configuration is adopted.

In addition to the state in which one node becomes redundant among the nodes having a redundant configuration as in the prior art, both nodes can be activated in response to a network failure. As a result, it becomes possible to more flexibly maintain the communication path between the rings.

The activated node sends a message packet for deleting the contents of the forwarding table registered by the learning bridge to each node in the ring, and the registration is deleted from the forwarding table according to the message contents. , Re-learning of the transfer table is performed earlier than the aging method.

It is also possible to shorten the time until re-learning by shortening the aging time instead of deleting the registration by the registration table deletion message of the transfer table.

In the present embodiment, the primary,
It is described that both the secondary nodes are notified by the mutual node MAC address setting, but for example, the MAC of the contact node paired with the contact node setting.
When the address is not set, an ARP (Address Resolution Pr) that transmits a contact node search packet on the ring and waits for a response from the partner node
It is also possible to adopt a method such as the auto protocol.
In this case, the MAC address of the paired contact node can be known by transmitting the broadcast packet in the ring and the paired contact node returning a response.

[0086]

As described above, according to the present invention, the first and second contact nodes for connecting the first and second packet rings are provided, and the first and second contact nodes are made redundant. As a result, it is possible to achieve a plurality of ring connections, and it is possible to perform high-speed switching even when the rings are connected by a plurality of nodes for redundancy.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a packet ring network according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a contact node of FIG.

FIG. 3 is a block diagram showing an example of a network configuration of RPR.

4 is a diagram showing a format of a ring packet used in the network configuration example shown in FIG.

5 is a flowchart showing an operation of the contact node of FIG. 1 when the device is activated.

6 is a diagram showing an example of transition of a secondary node from a standby state to an active state in the ring network shown in FIG.

FIG. 7 is a flowchart showing a process of activating a contact node when another node fails according to an embodiment of the present invention.

[Explanation of symbols]

11a-11c, 21a-21d nodes 12, 12a, 12b contact node 22a, 22b User terminal 101,102 ring 111,112 points of failure 121 Active node determination means 122 Active / Standby switching means 123,124 ring packet blocking means 201 Inner ring 202 outer ring

Claims (8)

[Claims] 1. A first and a second contact node connecting between a first and a second packet ring, wherein the first and the second contact nodes are provided.
Packet ring network characterized by redundant connection nodes. 2. A means for passing and blocking a packet between the first and second packet rings is provided as the first and second packets.
2. The packet ring network according to claim 1, wherein the packet ring network is included in each of the contact nodes, and the packet between the first and second packet rings is passed in one of the first and second contact nodes. 3. A means for controlling to permit passage of a packet between the first and second packet rings according to a failure state of the first and second packet rings. The packet ring network according to claim 1 or 2, wherein the packet ring network is included in each contact node, and the passage of the packet is simultaneously permitted to each of the first and second contact nodes. 4. A forwarding table for each of the nodes in the packet ring, the contact node of the first and second contact nodes being permitted to pass packets between the first and second packet rings. 4. The packet ring network according to claim 1, wherein the address of the other contact node is deleted from the packet ring network. 5. A connection method between packet ring networks, characterized in that the first and second packet rings are connected by redundant first and second connection nodes. 6. The connection between packet ring networks according to claim 5, wherein one of said first and second contact nodes passes a packet between said first and second packet rings. Method. 7. The connection method between packet ring networks according to claim 5 or 6, wherein passage of the packet is simultaneously permitted to each of the first and second contact nodes. 8. A forwarding table for each of the nodes in the packet ring, the contact node of the first and second contact nodes being permitted to pass a packet between the first and second packet rings. 8. The connection method between packet ring networks according to claim 5, wherein the address of the other contact node is deleted from.