JP5524934B2 - Recovery methods in communication networks - Google Patents

Description

  The present invention relates to a network recovery method in a connection type network. This has particular application in connection-oriented Ethernet networks with Multiprotocol Label Switching (MPLS).

  Network recovery methods generally detect a failure in one of the nodes in the network or an inter-node connection, and reroute (send along another path) traffic to bypass the failure, thereby causing a failure in the network. Configured to compensate. A network typically has a number of edge nodes where traffic can enter and exit the network and a number of intermediates where traffic can pass to travel from any one edge node to any other Node. Customer equipment configured to communicate over a network will typically communicate with one or more edge nodes. In the simplest type of approach, any customer device can only communicate with one edge node. Thus, any reroute performed by the recovery method cannot bypass an ingress node where traffic enters the network or an egress node where traffic exits the network. In some systems, dual parenting methods are known that allow customer equipment to communicate with two or more ingress or egress nodes. This has the advantage that the customer equipment can communicate over the network even when one of the edge nodes fails. However, there is currently no such dual parenting method available for MPLS or other connection-oriented networks. This means that the recovery method for an MPLS network relies on the diversity of paths between the same ingress and egress nodes, and thus the degree of recovery it can provide is limited. .

  The present invention is therefore preferably aimed at reducing, mitigating or removing at least one of the above-mentioned drawbacks, alone or in any combination.

  According to a first aspect of the present invention, there is provided a connection-type communication network including a plurality of nodes including a plurality of edge nodes, the network including a primary tunnel that connects a primary pair of the edge nodes; Configured to define a secondary tunnel that connects a secondary pair of the edge nodes that is different from the primary pair, and the network routes traffic when a failure is detected that affects the primary tunnel. A network is provided that is configured to be able to switch from a primary tunnel to the secondary tunnel.

  The first and second pairs of nodes may have a common node, but each has at least one node that is not common to both pairs.

  According to a second aspect of the present invention, there is provided a node for a connection-type network having a plurality of edge nodes, the node comprising: a primary tunnel connecting a primary pair of the edge nodes; and the primary pair Is configured to identify different secondary tunnels connecting the secondary pair of edge nodes, and the node can pass traffic from the primary tunnel to the secondary tunnel when a failure in the primary tunnel is detected. A node is provided that is configured to be switchable to a tunnel.

  Further features of the invention are set out in the claims as dependent claims.

  The present invention enables beneficial network recovery in an MPLS network or other connection-oriented network having a variety of nodes.

1 is a diagram illustrating an MPLS network according to an embodiment of the present invention. It is a state diagram which shows the operation | movement change of the network of FIG. FIG. 2 shows an MPLS network according to a further embodiment of the present invention. FIG. 2 shows an MPLS network according to a further embodiment of the present invention.

  Hereinafter, an example of a preferred embodiment of the present invention will be described with reference to the accompanying drawings.

  Referring to FIG. 1, a multiprotocol label switching (MPLS) network 10 according to an embodiment of the present invention includes a number of edge nodes 12, 14, 16, and 18 and a number of intermediate nodes 20 and 22. Here, various devices referred to by customer devices denoted by reference numerals 24 and 26 are configured to communicate via the network 10. Although only two customer equipment units 24, 26 are shown here for simplicity, it will be understood that in practice this number is generally much larger.

  As is conventional, each of the nodes 12, 14, 16, 18, 20, and 22 has one or more processors and associated memory and is configured to perform a number of functions related to the operation of the network. These functions are grouped into planes from an organizational perspective. The management plane has a network management system (NMS) that provides overall management of the network and provides an interface for the user to provide manual control of the network. The control plane uses a number of protocols and allows communication and forwarding (forwarding) plane settings between nodes. The transfer plane executes a transfer operation for transmitting data over the network. Control plane protocols that can include, for example, OSPF-TE and RSVP-TE are also configured to provide communication between edge nodes 12, 14, 16, 18 and optionally customer equipment 24, 26. ing. One of the main functions of the control plane protocol is to determine the path taken when all communications pass through the network.

  In order to allow two customer equipment (CE) units 24, 26 to communicate with each other, the network control plane protocol identifies one primary edge node 12 for one of the customer equipment units 24, and One primary edge node 12 is identified for the other customer equipment unit 26. For communication from the first unit 24 to the second unit 26, the first primary edge node 12 is defined as a primary ingress node and the second primary edge node 14 is defined as a primary egress node. . The control plane protocol also defines a number of communication tunnels that each provide communication from one of the edge nodes 12, 14, 16, 18 to the other nodes. Each of the tunnels is a path between two nodes capable of transmitting traffic in the network 10, so that traffic in the form of packets in the MPLS network is opaque to intermediate nodes. In an MPLS network, this is feasible because the intermediate node only reads the label of each packet and does not read the data in the packet. The ingress node 12 that directs traffic over the network identifies the egress node 14 that needs to transmit traffic to it and attaches a label to the information identifying the tunnel through which it is transmitted. Each tunnel is defined with a number of parameters, including two edge nodes to which the tunnel connects. Other defining parameters of the tunnel include, for example, the characteristics of traffic that can be carried, the quality of service (QoS) technique, the class of service associated with the MPLS tunnel, the maximum and average bandwidth, and the maximum burst size. Can be included, but in the present embodiment, all are included. Therefore, it is possible to have two or more tunnels between any pair of edge nodes and having various characteristics or parameters. Each tunnel may have only one label switch path (LSP) or multiple LSPs, but the best one depends on the characteristics and performance required by the client service , Will be selected at any given time. Customer traffic carried in the tunnel can be encapsulated in a pseudowire configured to support Ethernet services or any type of service allowed in the network Will understand.

  In communication over a network, a number of checks are performed by nodes defined by the control plane protocol to identify faults in the network. A number of fault detection methods are widely known, but will not be described in detail. A management information base (MIB) associated with the control plane and assisting the control plane is configured to store data regarding what faults have been detected in the network. In addition, all nodes in the network communicate with each other to check for failures in the network and communicate information about failures in the network, which affects which tunnels are available at any given time. To do. This type of check is performed when the CE unit 24 first transmits a transmission communication over the network, which will be described in detail later. This makes it possible to select a tunnel to be used depending on what kind of failure exists.

  Regardless of the type of communication between the CE unit 24 and the CE unit 26, a primary tunnel having one edge node 12 as an ingress node and one edge node 14 as an egress node is defined. Is done. In the present embodiment, the transmission CE unit 24 is linked to and communicable with not only the primary entry node 12 but also the secondary entry node 16, and the reception CE unit 26 is not only the primary exit node 14 but also the secondary exit node 18. It is linked to and can communicate. Accordingly, a primary tunnel AC is defined from the primary entry node 12 to the primary exit node 14. In addition, three secondary tunnels are also defined. That is, the secondary tunnel AD from the primary entry node 12 to the secondary exit node 18, the secondary tunnel BC from the secondary entry node 16 to the primary exit node 14, and the secondary tunnel from the secondary entry node 16 to the secondary exit node 18 BD. Thus, each tunnel exists between a pair of different nodes, and each tunnel shares one node with one of the other tunnels and the other node with another tunnel, but the rest Neither node shares with the tunnel.

  If the sending CE unit 24 wishes to initiate communication with the CE unit 26, it is arranged to direct communication traffic to the primary entry node 12. This traffic includes the identification information of the customer equipment unit 26 that is the transmission destination of communication, and information related to the characteristics of communication. Upon receiving this communication, the primary ingress node initiates transmission over the network to the primary egress node, triggers a number of communications between the edge nodes, and requires a separate tunnel from the primary tunnel used for communication. It is configured to check for any faults that would be. It will be appreciated that these communications between edge nodes can take a number of formats, but these should be sufficient to identify which tunnels are available and which are not available. Let's go.

  In this embodiment, the primary ingress node 12 is an active ingress node here, but transmits traffic from the CE unit 24 to the primary egress node 14. When the active egress node receives the traffic, it sends it to the receiving CE unit 26 to monitor the traffic. Appropriate means are used to check whether the LSP tunnel has successfully transferred traffic from the active ingress node to the active egress node using standard messages or dedicated protocols. As a result, the active ingress node 12 can determine whether or not a failure exists in the primary tunnel AC. If no failure is detected in the primary tunnel, traffic continues to be transmitted through the primary tunnel.

  In addition, when communication starts, the primary edge node and the secondary edge node are either the currently active tunnel, the secondary or standby tunnel, and either the primary tunnel or the secondary tunnel. A series of messages are configured to be exchanged so that each can determine if any failure exists. It will be appreciated that this message can take many forms. In this embodiment, a status message for verifying the integrity of the LSP is also transmitted to the standby ingress node 16. Appropriate signaling messages are also exchanged between the primary ingress node 12 and the secondary ingress node 16 to identify each state as an ingress node. This allows the active ingress node 12 to check whether the secondary ingress node 16 has experienced a failure. In order to be able to identify the state as the secondary leaving node, the active ingress node 12 also sends a check signal to the secondary leaving node 18. If necessary, the active ingress node 12 may check whether the standby egress node 18 is available using a keep-alive signaling message that is periodically exchanged between the two nodes. Is possible. In addition, the active ingress node 12 returns a confirmation signal to the transmitting CE unit 24 to confirm that it is receiving traffic and transmitting traffic to the receiving CE unit 26 over the network. Also configured to do. All these signals and checks can be repeated to continuously check the status of each tunnel during communication. These signals and checks can be exchanged between the primary edge node and the standby edge node in order to be able to check the state of the network.

  While the primary tunnel AC is functioning and continuing to provide the necessary communication, communication will continue on it. However, a failure is detected at either the primary edge node 12, 16 or at any intermediate node 20 that actually forms part of the primary tunnel AC that cannot be bypassed within the primary tunnel. If so, the network is configured to switch to one of the standby tunnels to allow communication to continue. The most appropriate tunnel is selected based on which node and LSP are functioning correctly. Each of the nodes 12, 14, 16, 18, 20, and 22 is configured to monitor and identify faults during communication, any of which may or may not require an active tunnel change. If a possible failure is identified, this is configured to communicate to each of the other edge nodes. In this way, all edge nodes are either edge nodes 12, 14, 16, 18 or intermediate nodes 20, 22 which can affect which tunnel is currently active and the selection of the active tunnel. It is possible to keep an accurate record of whether there are any faults in

  Referring to FIG. 2, the state of the network can be described as an active tunnel and how this changes due to the detection of various failures. When the ingress nodes 12 and 16 are indicated by A and B and the egress nodes 14 and 18 are indicated by C and D, the state of the network is a pair of actives, for example, AC when the primary tunnel is active. It can be indicated by a simple node. Thus, each state corresponds to a different active tunnel between different pairs of nodes. In FIG. 2, four states are shown, arrows between states indicate possible transitions between states, and each arrow label indicates a node with one letter or an LSP between two nodes with the appropriate two letters. Here, a failure causes a transition to occur. When communication begins, the system first transitions to state AC because AC is the primary or default tunnel. However, if a failure is then detected somewhere in the leaving node C or anywhere in the LSP AC between nodes A and C, the network switches to state AD and communication is sent to the standby leaving node D. Similarly, if a failure is detected at ingress node A, the network switches to state BD and traffic is communicated via tunnel BD. You will understand that there is some flexibility in what state the network will transition to when it detects a specific fault. Apart from the need to avoid all nodes and LSPs with known failures, this selection may further depend on other factors such as traffic conditions on the entire network.

  If the active ingress node 12 receives a message indicating a failure affecting the primary tunnel and is still functioning itself, the ingress node 12 can redirect the traffic to the appropriate secondary tunnel. Yes, the new active egress node can transmit traffic to the receiving CE 26. Thus, neither CE unit 24, 26 needs to perform the active part of the process, except that it is clearly necessary that the receiving CE is able to receive data from the secondary egress node. In the event of a failure that requires switching to the secondary entry node 16, communication is communicated within the secondary tunnel. The CE unit 24 needs to be able to transmit data not only to the primary entry node 12 but also to the secondary entry node 16. Optionally, external means can be used to communicate with the sending CE unit 24 to command the transmission of traffic to the secondary ingress node 16. This measure can be a dedicated message set or a dedicated extension to an existing protocol (eg, spanning tree) where the CE unit can direct traffic to the secondary ingress (without waiting for the learning process to occur). It is thought that the time required to recognize the necessity of transmission to the node is shortened. Accordingly, the CE unit 24 needs to be configured to switch from directing traffic to one of these to sending traffic to the other in response to a command from either ingress node 12,16. . However, in this embodiment, the transmitting CE unit 24 does not play an active role in determining the ingress node with which it should communicate. However, it is not necessary for the sending CE unit 24 to be completely unaware of failures that occur in the MPLS network.

  Referring to FIG. 3, in the second embodiment of the present invention, the basic configuration is the same as that of the first embodiment, and corresponding parts are indicated by the same reference numerals increased by 100. However, in this case, there is only one ingress node 124 that can communicate with the primary egress node 114 or the secondary egress node 118 via the primary tunnel 150 and the secondary tunnel 162, respectively.

  Referring to FIG. 4, in the third embodiment, the service provider network 210 has two MPLS networks 211 and 213 operated by different network operators. The first CE unit 224 communicates with the edge node 212 of the first network 211, and the second CE unit 226 communicates with the edge node 256 of the second network 213. Between these two edge nodes 212 and 256, a first tunnel passing through primary edge nodes 214 and 252 in the first and second networks is defined. A second tunnel is defined through secondary edge nodes 218, 254 in the first and second networks. Here, all of the edge nodes 212, 214, 218, 252, 254, 256 communicate with each other in the same manner as in the first embodiment of FIG. 1, thereby requiring transmission of traffic to the secondary tunnel. Any failure affecting the primary tunnel is identified and configured to keep the state of each of the nodes and each of the tunnels up to date as in the above-described embodiment.

Claims (10)

A connection-type communication network having a plurality of nodes including a plurality of edge nodes,
The network is configured to define a primary tunnel connecting the primary pair of edge nodes and a secondary tunnel connecting the secondary pair of edge nodes different from the primary pair;
The primary tunnel and the secondary tunnel have different first ingress nodes and second ingress nodes,
The network may be configured to switch the first ingress node or the second ingress node to switch traffic from the primary tunnel to the secondary tunnel when a fault affecting the primary tunnel is detected . A network that is configured to instruct customer devices from either .   The network according to claim 1, wherein the primary tunnel and the secondary tunnel have different egress nodes.   The at least two of the edge nodes are configured to communicate with each other using fault identification communication to identify related faults that require switching to the secondary tunnel. 2. The network according to 2.   The network according to claim 3, wherein the failure identification communication is performed between the ingress node of the primary tunnel and the egress node of the primary tunnel.   The network according to claim 3 or 4, wherein the failure identification communication is performed between a node of the primary tunnel and a node of the secondary tunnel.   The network according to claim 3, wherein the failure identification communication is performed between the entry node of the primary tunnel and the entry node of the secondary tunnel.   The said failure identification communication is performed between the said leaving node of the said primary tunnel, and the said leaving node of the said secondary tunnel, Any one of Claim 3 to 6 dependent on Claim 2 characterized by the above-mentioned. The network described in the section.   A plurality of entry nodes and a plurality of exit nodes;
  The network according to claim 1, wherein a tunnel is defined between each of the plurality of ingress nodes and each of the plurality of egress nodes.   A node for a connection type network having a plurality of edge nodes,
  The network is configured to define a primary tunnel connecting the primary pair of the edge nodes and a secondary tunnel connecting the secondary pair of the edge nodes different from the primary pair, the primary tunnel And the secondary tunnel has different ingress nodes,
  The node is one of the different ingress nodes and directs customer equipment to switch traffic from the primary tunnel to the secondary tunnel if a fault affecting the primary tunnel is detected Configured to
  A node characterized by that.   10. The edge node configured to communicate with other edge nodes in the network to identify a failure in the network that requires switching the traffic to the secondary tunnel. The listed node.

Images