Solution to routing loop problem of Distance Vector Routing Protocol

In summary, there are six solutions to the routing loop problem in distance vector routing protocol:

1. Define the maximum value;

2. horizontal segmentation technology;

3. Route poisoning;

4. Reverse route poisoning;

5. Control the update time;

6. Trigger the update.

The following describes the implementation principles of various solutions:

1. Define the maximum value:

Distance Vector routing can be corrected by TTL in the IP Address Header, but the routing loop problem may require infinite count first. To avoid this latency problem, the distance vector protocol defines a maximum value, which refers to the maximum measurement value (for example, the maximum rip Protocol value is 16), such as the number of hops. That is to say, the route update information can be sent 15 times to the router in the inaccessible network. Once the maximum value is 16, the network is deemed to be inaccessible and there is a fault, no route updates from accessing the network will be accepted.

2. horizontal segmentation:

A method to eliminate routing loops and accelerate network convergence is implemented through a technology called horizontal segmentation. The rule is not to re-send route update information to the original route update direction (one-way update, one-way feedback ). For example, if three routers ABC and B learn from C the path to 10.4.0.0, they no longer declare to C that they can access the path of 10.4.0.0 through C, after learning the network path to 10.4.0.0 from B, A does not declare it to B. However, once the network 10.4.0.0 fails, C will send the route update information that cannot be achieved by the network to A and B, but will not learn the error message sent by A and B that can reach 10.4.0.0.

3. Route poisoning (also known as route poisoning ):

Defining the maximum value solves the routing loop problem to a certain extent, but it is not thorough. We can see that the routing loop still exists before it reaches the maximum value. Therefore, routing poisoning can completely solve this problem. The principle is as follows: assume there are three routers ABC. When the network 10.4.0.0 fails and cannot be accessed, router C sends the route update information to the neighbor route, and mark the measurement value as infinity to inform them that the network 10.4.0.0 cannot be reached. After receiving the poison message, vrob B marks the route table entry of the link as Infinity, indicating that the path has failed, then, the customer notifies the neighbor router A that each router has been poisoned, and the network 10.4.0.0 has failed and no longer receives updates. This avoids the routing loop.

4. Reverse poisoning (also known as drug reversal ):

In combination with the preceding example, when router B sees that the measurement value reaching network 10.4.0.0 is infinite, it sends an update message called drug reversal to router C, indicating that the network 10.4.0.0 cannot be reached, this is a feature that goes beyond the horizontal split. This ensures that all routers receive toxic routing information.

5. Control the update time (that is, suppress the timer ):

The suppression timer is used to prevent messages that are periodically updated from resetting a broken route in an inappropriate time. The suppression timer tells the router to temporarily retain any changes that may affect the route for a period of time. The suppression time is usually longer than the time for updating the information to be sent to the entire network. When a router receives an update from a neighbor that is inaccessible to the previously accessed network, it marks the route as inaccessible and starts a suppression timer, if you receive the update information sent from the neighbor again, which contains a route with a better measurement value than the original path, mark it as accessible and cancel the suppression timer. If the update information received from different neighbors before the suppression timer times out contains a worse metric value than the previous one, the update will be ignored, in this way, you can spend more time spreading updates throughout the network.

6. Trigger update:

Under normal circumstances, the router regularly sends the route table to the neighbor router. When an update is triggered, the route update information is immediately sent to respond to certain changes. The router that detects a network failure immediately sends an update message to the neighbor router, and generates the neighbor router that triggers the update notification in turn, make the router on the entire network receive the update information in the shortest time, so as to quickly understand the changes of the entire network. However, this problem also exists. data packets that may contain updates may be lost or damaged by links in some networks. Other routers fail to receive the update trigger in time, therefore, an update triggered by suppression is generated. Once a route is invalid, routes with the same or worse metric values reach the same destination during the suppression time are ignored, in this way, the update trigger takes some time to spread throughout the network. This prevents the damaged route from being re-inserted into the neighbor that has received the update trigger and solves the routing loop problem.

The following message is displayed for Route poisoning and reverse poisoning:

Route poisoning: When the routing information fails in the routing table, the measurement value is changed to infinity instead of deleting the route information from the routing table immediately. (This sentence should be understood. For example, in the RIP Protocol, the measurement value is changed to 16, which means the route is not reachable.) then, the information is published, in this way, the adjacent router will know that this route is invalid .....

The concept of toxicity inversion is different from that of Route poisoning. It refers to the router that receives the route poisoning message and does not follow the horizontal segmentation principle to forward the poisoning message to all adjacent routers, it also includes the source router that sends the poisoning information, that is, notifying the adjacent router that the route information has failed. The main purpose is to achieve rapid convergence.