KR20220028319A - Microloop avoidance method - Google Patents

Abstract

A micro-loop avoidance method of the present invention is a method of preventing micro-loops between router nodes caused by topology changes caused by network failures or recovery. In the present invention, a method of sequentially applying the routing table reflection reflecting the topology change according to the metric value is used, and tunneling is used to prevent micro-loops when a general router which does not apply sequentially is included in a path.

Description

Microloop avoidance method {MICROLOOP AVOIDANCE METHOD}

The present invention relates to a network routing technology, and more particularly, to a method of preventing a microloop caused by a link failure or the like.

A routing protocol is a communication protocol that defines a communication method between routers. Although many kinds of routing protocols are defined in RFC documents, the following three main classes are mainly used in IP networks. Open Shortest Path First (OSPF) and Intermediate System to Intermediate System (IS??IS) for internal gateway routing through link state routing protocols, and Interior Gateway Routing Protocol (IGRP) for internal gateway routing through path vector or distance vector protocols ), Enhanced Interior Gateway Routing Protocol (EIGRP), and Border Gateway Protocol (BGP) for external gateway routing are major classes.

For example, OSPF is a dynamic routing protocol and is a representative link state routing protocol. The optimal path is selected by monitoring the status of the connected link.

The Fast Reroute (FRR) technique is a technique for recovering traffic with minimal packet loss in the event of a network failure. FRR calculates and sets an alternate path to be used in the event of a failure, that is, a backup path in advance. In case of a failure, traffic is detoured to the backup path.

Even with FRR, a microloop may occur between router nodes due to the time difference of installing a new route in the FIB between each router node constituting the routing domain.

Registered Patent No. 10??2123689 discloses a metric delay, which is a method of avoiding a microloop caused by a difference in FIB installation time. However, the method disclosed in the registered patent works effectively when all router nodes in the network support the metric delay, and when a general router node that does not support the metric delay is included in the network, the metric delay method prevents micro loops. Difficult situations may arise.

Republic of Korea Patent Registration No. 10??2123689

It is an object of the present invention to provide a microloop prevention method that can be applied even if a general router node that does not support a method for preventing a microloop caused by a difference in installation time in the FIB of a router node is included in the network.

According to one aspect of the invention, a microloop avoidance method due to a topology change occurring in a network including both a router node supporting metric delay and a general router node not supporting metric delay is provided in the network. A metric delay advertisement stage in which router nodes advertise whether or not the metric delay is supported, a backup path activation stage in which a router node in which a topology change event occurs activates a preset backup path, and a router node in which the topology change event occurs is another router node in the network Topology change notification step of notifying topology change notification information with Tunneling setup stage in which one router node establishes tunneling to the closest metric delay-supporting router node to the node where the topology change event occurred, and the router node recognizing the topology change has a different delay time depending on the calculated router metric value It includes a shortest path reflection step in which the selected shortest path is reflected in the routing table after waiting for , and a tunneling release step in which a router node that has set up tunneling releases the tunneling.

According to another aspect of the invention, the delay time the router node waits for in the shortest path reflection step may be calculated as a product of a preset delay constant value and a router metric value.

According to another aspect of the invention, the preset delay constant value in the shortest path reflection step may be set to a different value in at least one router node.

According to another aspect of the present invention, a microloop avoidance method includes the steps of notifying, by a router node, a set delay constant value to another router node, before a topology change event occurs, and receiving the set delay constant value of another router node. The method may further include determining, by the router node, the delay constant value as a larger value compared with the set delay constant value.

According to the microloop avoidance method using the metric delay of the present invention, the router node calculates the shortest path for the topology changed due to the failure occurring in the network, and before reflecting the calculated shortest path in the FIB of the router node, the router node performs the metric After waiting for the delay time calculated using the value, the calculated shortest path is reflected in the FIB to prevent the occurrence of microloop and minimize packet loss. The purpose of this is to provide a method to prevent micro-loop through tunneling when a general router node is included.

1 is a diagram illustrating the occurrence of a microloop due to a shortest path change due to occurrence of a link failure.
2 is a diagram illustrating the occurrence of a microloop due to a shortest path change due to link failure recovery.
3 is a diagram illustrating activating a backup path and detouring traffic in case of a link failure according to an exemplary embodiment.
4 is a diagram illustrating that the shortest path is changed without generating a microloop when a link fails in a network including only router nodes supporting metric delay.
5 is a diagram illustrating that the shortest path is changed without generating a microloop when a link is restored in a network composed only of router nodes supporting metric delay.
6 is a diagram illustrating a micro-loop occurring in the event of a link failure due to a general router node that does not support metric delay.
7 is a diagram illustrating prevention of microloop generation in case of link failure in a network including a general router node that does not support metric delay.
8 is a flowchart illustrating a micro-loop avoidance method in a network including a general router node that does not support metric delay.

The foregoing and additional aspects are embodied through embodiments described with reference to the accompanying drawings. It is understood that various combinations of elements in each embodiment are possible within the embodiments as long as there is no contradiction between them or other mentions. Each block in the block diagram may in some cases represent a physical part, but in other cases may be a part of the function of one physical part or a logical representation of a function across a plurality of physical parts. Sometimes a block or part of an entity may be a set of program instructions. All or a part of these blocks may be implemented by hardware, software, or a combination thereof.

1 is a diagram illustrating the occurrence of a microloop due to a shortest path change due to occurrence of a link failure. 1 exemplarily shows a ring topology for a router domain composed of nine router nodes. The router node shown in FIG. 1 is a device that designates a route for the transmission of IP packets, ie, traffic, in a network, and finds and delivers the fastest and most efficient route for IP packets from a source to a destination. In FIG. 1, a router node 9 is a gateway to another router domain. For the sake of explanation, it is assumed that the link costs between each router node are all the same. For example, when Open Shortest Path First (OSPF) is used as the routing protocol in FIG. 1 , link costs between router nodes are typically defined according to link speeds, so it is assumed that all links are set to the same speed. A packet originating from router node 4 to be forwarded to an external router is transmitted to its destination through router node 3/router node 2/router node 1/router node 9 along the shortest path. Similarly, a packet originating from router node 5 is transmitted to its destination through router node 6/router node 7/router node 8/router node 9 along the shortest path.

At this time, as shown in FIG. 1, when the link between router node 1 and router node 9 is down, a topology change event (link change event) occurs in router node 1 and router node 9, and router node 1 and router node 9 transmits Topology Change Notification information to other router nodes in the same routing area. For example, when the routing protocol is IS??IS, it is transmitted to other router nodes through LSP (Link State Packet), and when the routing protocol is OSPF, it is transmitted to other router nodes through LSA (Link Status Advertisement).

The router node where the topology change event occurred or the router node that received the topology change notification information calculates SPF (Shortest Path First) and selects a new shortest path. The new shortest path selected is installed in the Forwarding Information Base (FIB), which is a routing table located in the data plane that forwards packets from the router, that is, reflected. At this time, the FIB installation time is different between each router node, and a micro-loop occurs between adjacent router nodes. In the example shown in Fig. 1, between router node 2 and router node 3, router node 2 first installs a new route in the FIB and forwards the packet to router node 3, but router node 3 still forwards the new route to the FIB. If the installation fails and router node 3 sends the packet back to router node 2 again, a micro-loop temporarily occurs between router node 2 and router node 3, which causes the packet until router node 3 installs a new route into the FIB. loss occurs. A microloop may occur between router node 1 and router node 2, between router node 2 and router node 3, and between router node 3 and router node 4 in FIG. 1 .

At this time, the router node 5/6/7/8/9 does not change the shortest path due to the topology change, so that the next hop on the alternate SPF path from that shown in FIG. 1 is not changed.

2 is a diagram illustrating the occurrence of a microloop due to a shortest path change due to link failure recovery. As shown in FIG. 2 , a microloop may occur even when a link failure is restored. FIG. 2 shows that the shortest path is changed due to a link failure between the router node 1 and the router node 9, and then the shortest path is changed again after the link failure is restored. As in the case of FIG. 1 , a topology change event occurs in router node 1 and router node 9, and router node 1 and router node 9 transmit topology change notification information to other router nodes in the same area. Each router node selects a new shortest path by calculating SPF, and installs the new shortest path selected in the FIB. At this time, the FIB installation time is different between each router node, and a micro-loop occurs between adjacent router nodes. In the example shown in Fig. 2, between router node 2 and router node 3, router node 3 first installs a new route in the FIB and forwards the packet to node 2, but router node 2 has not yet installed a new route in the FIB. Still, if router node 2 sends the packet back to router node 3, a micro-loop temporarily occurs between router node 2 and router node 3, which prevents packet loss until router node 2 installs a new route into the FIB. Occurs. A microloop may occur between router node 1 and router node 2, between router node 2 and router node 3, and between router node 3 and router node 4 in FIG. 2 .

3 is a diagram illustrating activating a backup path and detouring traffic in case of a link failure according to an embodiment, and FIG. 4 is a shortest path change without a microloop in case of a link failure in a network consisting only of router nodes supporting metric delay It shows what will be That is, FIGS. 3 and 4 show a process in which the next hop of router nodes is changed by the metric delay method.

The metric delay is a name defined for the convenience of description of the method for avoiding micro loops of the present invention. A router node supporting metric delay is a router capable of performing the procedure of the microloop avoidance method of the present invention and is expressed as a metric delay supporting router node, and a router node not supporting metric delay is expressed as a general router node. All router nodes (except external routers) shown in FIGS. 3 and 4 are metric delay support router nodes. A metric delay during link switching will be described with reference to FIGS. 3 and 4 .

As shown in FIG. 3 , when a link is switched between router node 1 and router node 9, router node 1 applies Fast Reroute (FRR) to activate a preset backup path and detour packets for quick failure recovery. In the ring topology of FIG. 3 , a backup tunnel is made to the PQ node (router node 5) using the Remote Loop Free Alternative (RLFA) protocol to prevent packet loss. At this time, router node 2, router node 3, and router node 4 still transmit the packet through the shortest path before the link changeover event, and router node 1 transmits the packet using the backup tunnel. That is, the microloop avoidance method according to an embodiment minimizes packet loss by using the FRR protocol when a link changeover event occurs.

Router node 1 and router node 9 forward topology change notification (TCN) information to other router nodes in the same routing area. For example, when the routing protocol is IS??IS, it is transmitted to other router nodes through LSP (Link State Packet), and when the routing protocol is OSPF, it is transmitted to other router nodes through LSA (Link Status Advertisement).

In the example of the metric delay shown in FIG. 4, each router node that has received TCN information from router node 1 and router node 9, which is a router node where a link change event has occurred in the event of a link failure, starts an SPF timer, and the SPF timer is terminated ( Expire), SPF is performed to re-elect the shortest path. The shortest path to the newly elected destination has new metric information. In the example of FIG. 4, assuming that the link sections between each router node all have the same metric of 10, in the case of a route to router node 9, the metric increases by 10 every time it passes a hop to router node 9. Therefore, for router node 4, the metric is 50, for router node 3, the metric is 60, for router node 2, the metric is 70, and for router node 1, the metric is 80.

When SPF calculation is completed, each router node applies new route information directly to FIB. However, in metric delay, the delay time is calculated using the metric value instead of directly applied to the FIB, and FIB installation is delayed by that time to generate a micro loop. can be minimized There is no restriction on the formula for calculating the delay time using the metric value, but each router node must be able to install route information into the FIB at different times according to the metric value. As an example, the formula for calculating the delay time converts a result of multiplying a metric value by a preset delay constant (Max convergence value, 100 in the example of FIG. 4 ) into a time in ms, and delays it by that time. That is, in the case of router node 4, it is 100*50 (Metric), delaying 5000ms, that is, 5 seconds, and then installing the FIB. In the case of , the FIB is installed after a delay of 8 seconds. In the example of FIG. 4 , when new routes are sequentially installed in the FIB in the order of router node 4/3/2/1 and router node 3 installs the FIB, since router node 4 has already installed the FIB, router node 3 and router A microloop does not occur between Node 4. Similarly, a micro-loop does not occur between router node 2 and router node 3 and between router node 1 and router node 2. In case of router node 1, the backup path is also disabled along with FIB installation.

5 is a diagram illustrating that the shortest path is changed without generating a microloop when a link is restored in a network composed only of router nodes supporting metric delay. A metric delay in link recovery will be described with reference to FIG. 5 . During link recovery, each router node that receives TCN information from router node 1 and router node 9, which is the router node where the link change event occurred, starts an SPF timer. When the SPF timer expires, SPF is performed to find the shortest path re-elect The shortest path to the newly elected destination has new metric information. In the example of Fig. 5, assuming that the link sections between each router node all have the same metric of 10, in the case of a route to router node 9, the metric increases by 10 every time it passes a hop to router node 9. Therefore, in the case of router node 1, the metric is 10, in the case of router node 2, the metric is 20, in the case of router node 3, the metric is 30, and in the case of router node 4, the metric is 40.

When the SPF calculation is completed, each router node calculates the delay time using the metric value and delays the FIB installation time by that time to minimize the occurrence of micro loops. There is no restriction on the formula for calculating the delay time using the metric value, but each router node must be able to install route information into the FIB at different times according to the metric value. As an example, the formula for calculating the delay time converts a result of multiplying a metric value by a preset delay constant (Max convergence value, 100 in FIG. 5) into a time in ms, and delays the time by that time. That is, in the case of router node 4, it is 100*40 (Metric), delaying 4000ms, that is, 4 seconds, and then installing the FIB. In the case of , the FIB is installed after a delay of 1 second. In the example of FIG. 5, when new routes are sequentially installed in the FIB in the order of router nodes 1/2/3/4 and router node 3 installs the FIB, since router node 2 has already installed the FIB, router node 3 and router A microloop does not occur between Node2. Similarly, a micro-loop does not occur between router node 4 and router node 3 and between router node 1 and router node 2.

FIG. 6 shows microloops occurring when a link failure occurs due to a general router node that does not support metric delay, and FIG. 7 is a microloop that occurs when a link failure occurs in a network including a general router node that does not support metric delay. shown to be prevented. In the examples of FIGS. 6 and 7 , router node 3 is a general router that does not support metric delay.

As shown in FIGS. 6 and 7, when a link is switched between router node 1 and router node 9, router node 1 applies FRR to activate a preset backup path and detour packets for fast failure recovery. Router node 1 and router node 9 forward topology change notification (TCN) information to other router nodes in the same routing area. Each router node that receives TCN information from router node 1 and router node 9, which is the router node where the link change event occurred, runs an SPF timer. When the SPF timer expires, SPF is performed to re-select the shortest path. The shortest path to the newly elected destination has new metric information. In the example of FIG. 6, assuming that the link sections between each router node all have the same metric of 10, in the case of a route to router node 9, the metric increases by 10 every time it passes a hop to router node 9. Therefore, for router node 4, the metric is 50, for router node 3, the metric is 60, for router node 2, the metric is 70, and for router node 1, the metric is 80.

When the SPF calculation is completed, router node 3, which is a general router, immediately applies the new route information to the FIB, but other router nodes supporting metric delay do not apply the new route information directly to the FIB, but calculate the delay time using the metric value. Delay FIB installation by that time. As an example, the formula for calculating the delay time converts a result of multiplying a preset delay constant (Max convergence value, 100 in the example of FIG. 6 ) by a metric value into a time in ms, and delays the time by that time. That is, in the case of router node 4, the FIB is installed after delaying 5000ms, that is, 5 seconds by 100*50 (Metric). That is, the next hop of the packet that needs to be forwarded to the external router for 5 seconds is still router node 3 in the case of router node 4. However, since the next hop of the packet to be forwarded to the external router of router node 3 has already been changed to router node 4, a micro-loop occurs for 5 seconds even though router node 4 supports the metric delay. This is because router node 3 does not support metric delay, and this situation can occur frequently.

As shown in FIG. 7 , router node 2 and router node 4 know that a router node that does not support metric delay exists on an existing path. All router nodes on the network advertise LSP (for IS??IS)/LSA (for OSPF). At this time, routers supporting metric delay advertise by including metric delay capability in this LSP/LSA. However, normal routers that do not support metric delay cannot advertise Metric Delay Capability, so other router nodes that support metric delay can identify the general router.

The router node 2 and the router node 4 of FIG. 7 establish tunneling with the router node 1 which is the closest metric delay support router to the node where the topology change event occurred. Router Node 2 and Router Node 4 encapsulate the packet destined for the external router and send it to the router node 1 as the destination. Therefore, even if router node 3 forwards a packet destined for an external router to router node 4, which is the next hop, router node 4 encapsulates the packet, adds an address so that it is directed to router node 1, and transmits it back to router node 3, is forwarded to router node 2, which is the next hop of the packet destined for router node 1, and router node 2 forwards it to router node 1. Router node 1 may delete the tunneling address of the encapsulated packet and forward the actual packet to the next hop, router node 5, so that the packet is forwarded to an external router. As such, since router node 3 does not support metric delay, packets can be forwarded without microloop even if the metric delay clause is not followed.

8 is a flowchart illustrating a micro-loop avoidance method in a network including a general router node that does not support metric delay. A micro-loop avoidance method according to an aspect is a method of avoiding a micro-loop due to a topology change occurring in a network including both a router node supporting metric delay and a general router node not supporting metric delay. It includes a delay advertisement step, a backup path activation step, a topology change notification step, a shortest path selection step, a tunneling setup step, a shortest path reflection step, and a tunneling release step.

In a network including multiple router nodes, that is, when a link failure between router nodes occurs in a routing domain, a topology change event occurs. A router node in which a topology change event occurs is called a Point of Local Repair (PLR).

The metric delay advertisement step is a step in which router nodes in the network advertise whether or not the metric delay is supported. All router nodes on the network advertise LSP (for IS??IS)/LSA (for OSPF). At this time, routers supporting metric delay advertise by including metric delay capability in this LSP/LSA. However, since a general router that does not support metric delay advertises an LSP/LSA that does not include metric delay capability, other router nodes that support metric delay can recognize the general router.

The backup path activation step is a step in which the router node where the topology change event occurred, that is, the PLR node, activates a preset backup path. The preset backup path is a path that bypasses packets to minimize traffic loss in case of link switching. When all router nodes on the network complete the SPF path calculation, they are added to the FIB in preparation for link breakage with neighboring nodes. Even if the backup path is installed in the FIB, the router node forwards traffic using the primary path if there is no failure in the primary path. send traffic. A technique for detouring packets using such a backup path is called FRR. The backup path may be an alternative next??hop different from the existing path, or may be an established backup tunnel.

The topology change notification step is a step in which the router node in which the topology change event occurs notifies the topology change notification information to all other router nodes in the network, that is, the routing domain.

In the shortest path selection step, each router node that detects a link failure to the PLR node or receives topology change notification information and recognizes the topology change waits for a certain amount of time, that is, the SPF timer time, and then when the SPF timer expires, the router metric to the destination is retrieved. This is the step of re-calculating and selecting the shortest path.

The tunneling setup step is a step in which the router node, recognizing the topology change, establishes tunneling with the metric delay support router node closest to the node where the topology change event occurred. A router node that establishes tunneling is a metric delay support router node. In the example of FIG. 7 , router node 3 does not establish tunneling, and router node 2 and router node 4 configure tunneling. When tunneling is established, if the destination of the packet that encapsulates the existing packet is the address of the other router node for which tunneling is set, and the corresponding packet is transmitted, the packet can be transmitted without occurrence of a microloop.

In the shortest path reflection step, each router node recognizing the topology change waits for the calculated delay time according to the calculated router metric value, and then reflects the selected shortest path in the routing table. At this time, the delay time calculated at each router node is calculated differently according to the metric value for each router.

The tunneling release step is a step in which the router node that has set up tunneling releases the tunneling.

The de-tunneling step and the shortest path reflection step are performed substantially simultaneously. Depending on the aspect, the tunneling release step may be performed first, and the shortest path reflecting step may be performed almost simultaneously.

Referring to FIG. 8 , when router nodes operate as IGP routers, LSP/LSA advertisements are performed, and at this time, metric delay capability is advertised (S1000). When the router node detects a link failure connected to its interface (S1020), it activates a preset backup path according to the FRR protocol to detour traffic to the corresponding path (S1040). The corresponding router node notifies the topology change notification information to all other router nodes in the same routing domain (S1060), and each router node selects the shortest path by performing SPF calculation again (S1080). At this time, each router node has a new metric value for the new shortest path. Each router node calculates a delay time using the new router metric value (S1100). If there is a general router node that does not support the metric delay among routers whose next hop is changed due to a topology change, the router node supporting the metric delay establishes tunneling with the router node closest to the node where the topology change event occurred (S1120). After waiting for the calculated delay time (S1140), the selected shortest path is reflected in the FIB (S1160). The router node that has set up tunneling releases tunneling (S1180).

According to another aspect of the invention, the delay time that each router node waits in the shortest path reflection step may be calculated as a product of a preset delay constant value and a router metric value.

According to another aspect of the invention, the change of topology occurs in case of recovery from link failure as well as link failure. If the topology change event is caused by the recovery of a link failure, the backup path in the backup path activation phase is the current transmission path before the link failure is restored. Therefore, in the case of link recovery, since the original path (the shortest path newly elected due to link change) already exists, the original path is the backup path until the new shortest path is installed in the FIB. Therefore, in this case, there is no need to separately activate the backup path.

According to another aspect of the invention, each router node in the shortest path reflection step may set a value of the delay constant differently. That is, the delay constant value preset in each router node may be set to a different value in at least one router node. The delay constant value can be set by the operator through CLI (Command Line Interface). In this case, the operator can set different settings for each router node in consideration of the number of routing entries that must be changed due to a topology change, CPU performance of the router node, and traffic throughput of the router node.

According to another aspect of the invention, when the delay constant value is set differently for each router node, the microloop avoidance method includes the steps of notifying the delay constant value set by each router node in advance to another router node before the topology change event occurs; The method may further include determining, by the router node, having received the set delay constant value of the router node, the delay constant value as a large value compared with the set delay constant value. The operator sets the delay constant value differently for each router node, and each router node can share the value through IGP. For example, in the case of a router using the OSPF routing protocol, RFC 4970, and in the case of a router using the ISIS routing protocol, RFC 4971 defines the Router Capability TLV, where the Originator Router advertises its properties, and defines a new Sub TLV. to share the value of the delay constant.

According to an embodiment, a router node using a method for avoiding micro-loop due to topology change upon network failure or recovery activates a preset backup path when a topology change event occurs, and notifies other router nodes of the network topology change. When a topology change event occurs or topology change notification information is received from other router nodes to recognize the topology change, it waits for a certain period of time and recalculates the router metric to the destination to select the shortest route and recognizes the topology change. The router node sets up tunneling to the metric delay support router node closest to the node where the topology change event occurred, calculates the delay time using the calculated router metric value, and each router node uses the calculated delay time. After waiting, the selected shortest path is reflected in the routing table, and after reflection, the configured tunneling can be released.

According to another embodiment, the delay time that the router node waits to reflect the selected shortest path in the routing table may be calculated as a product of a preset delay constant value and a router metric value.

According to another embodiment, the preset delay constant value may be set to a value different from that of other router nodes.

According to another embodiment, each router node may set a different delay constant value. In this case, the set delay constant value is notified to other router nodes in advance, and the set delay constant value is received when the other router node receives the set delay constant value. It is possible to determine the value of the delay constant as a large value compared to the constant value.

Although the present invention has been described above with reference to the accompanying drawings, the present invention is not limited thereto, and it should be construed to encompass various modifications that can be apparent from those skilled in the art. The claims are intended to cover such variations.

Claims (8)

A method for avoiding microloop due to topology change occurring in a network including both a router node supporting metric delay and a general router node not supporting metric delay, the method comprising:
a metric delay advertisement step in which router nodes in the network advertise whether or not the metric delay is supported;
a backup path activation step in which a router node in which a topology change event occurs activates a preset backup path;
a topology change notification step in which the router node in which the topology change event has occurred notifies the topology change information to other router nodes in the network;
a shortest path selection step in which the router node, recognizing the topology change, selects the shortest path by re-calculating the router metric to the destination after waiting for a certain period of time;
a tunneling setting step of establishing tunneling with a router node closest to the node in which the topology change event occurred among the router nodes recognizing the topology change;
a shortest path reflection step in which the router node recognizing the topology change calculates a delay time using the calculated router metric value, and the router node waits for the calculated delay time and then reflects the selected shortest path in the routing table; and
a tunneling release step in which a router node that has set up tunneling releases the tunneling;
including,
In the tunneling setup phase, the router node that establishes the tunnel is a metric delay support router node, a micro-loop avoidance method.
The method of claim 1,
A microloop avoidance method in which the delay time each router node waits in the shortest path reflection step is calculated as the product of the preset delay constant value and the router metric value.
3. The method of claim 2,
A microloop avoidance method in which a preset delay constant value in the shortest path reflection step is set to a different value in at least one router node.
4. The method of claim 3, wherein the method
before the topology change event occurs, each router node notifying the set delay constant value to other router nodes; and
determining, by a router node having received a set delay constant value of another router node, a larger delay constant value by comparing it with the set delay constant value;
A micro-loop avoidance method further comprising a.
A router node supporting a metric delay that avoids a micro loop due to a topology change in the event of a network failure or recovery, comprising:
When a topology change event occurs, a preset backup path is activated, and the topology change information is notified to other router nodes in the network.
When the topology change is recognized, the shortest route is selected by recalculating the router metrics to the destination after waiting for a certain period of time. ), calculates the delay time using the calculated router metric value, each router node waits for the calculated delay time, reflects the selected shortest path in the routing table, and releases the tunneling set after reflection. router node.
6. The method of claim 5,
The waiting delay time is calculated as the product of the preset delay constant value and the router metric value.
7. The method of claim 6,
A router node whose preset delay constant value is set to a value different from that of other router nodes.
8. The method of claim 7,
A router node that notifies the set delay constant value to other router nodes before the topology change event occurs, and determines the delay constant value with a larger value compared with the set delay constant value when receiving the set delay constant value of other router nodes.

Images