Shulou(Shulou.com)06/02 Report--

The following is excerpted from the complete manual of Cisco router configuration and management (second edition), one of the newly listed "four King Kong" books. (the other three books are Cisco switch configuration and management complete manual (second edition), H3C switch configuration and management complete manual (second edition) and H3C router configuration and management complete manual (second edition). At present, these four new books are on sale in all online stores: http://item.jd.com/11273171.html

8.1.3 RIP routing update mechanism

RIP protocol has two update mechanisms: one is to update periodically, and the other is to trigger updates. "periodic update" is to send RIP route advertisements periodically according to the set update timer. The advertisement message carries all the RIP routing information in the local router except the RIP routes suppressed by the split horizon mechanism. On the other hand, "triggered update" is a RIP route advertisement sent by a RIP router only when there is a change in the routing table entry, and only carries the route information with changes in the local routing table. As soon as the RIP router senses a change in the network, it sends update messages as soon as possible or even immediately, without waiting for the update cycle to end. "counting to infinity" can be greatly prevented as long as the update is triggered fast enough, but it is still possible.

Whether you update periodically or trigger updates, the update rules for RIP routes are as follows:

If an updated routing table entry is not in the routing table, the routing table entry is added directly to the routing table

If there is already a routing table entry for the same destination network in the routing table and the source port is the same, then the routing table is updated unconditionally according to the latest routing information.

L if there is already a routing table entry for the same destination network in the routing table, but the source port is different, compare their metrics and take the lower metric as your own routing table entry

If there is already a routing table entry for the same destination network in the routing table, and the metric is equal, the original routing table entry is retained.

The following is mainly about the periodic update mechanism of RIP routing.

1.RIP routing periodic update mechanism

The RIP router always sends a routing update packet to the neighboring router in RIP broadcast reply over the UDP port 520s (this is the default value, which can be modified, and may deviate from the setting). The packet includes the complete routing table on this router (except for the routing table entries suppressed by the split horizon mechanism) to provide routing updates to neighboring routers. At the same time, it is used to prove its existence to neighboring routers. The routing table of RIP mainly includes "destination network", "next-hop address" and "distance", as shown in figure 8-3.

If a router does not receive a routing update from a neighbor router within 180 seconds (which is also the default value and can be modified), the router marks the neighbor router as unreachable, leaving the neighbor router in a hold-down cycle. When a router is in a hold-down period, it is still used to forward packets, but other routers in the network do not learn routing information to the network to which the router is connected, unless it is a better route to the network to which the router is connected. for example, if it was originally 3 hops, you learned a 2-hop routing information during the suppression period. However, after the suppression period, even poor routing information is accepted.

If the routing update for this router has not been received within a continuous period of 240 seconds (which is also the default value and can be modified), the local router deletes the routing table entry associated with the neighbor router in the routing table.

Thus it can be seen that this routing update not only affects the update of the most table and all the packets that need to arrive on the router in the whole RIP network, or the routing of packets passing through the router, but also affects whether other neighboring routers exist. Just imagine, if a packet is to be sent to a host on a network connected to a certain RIP router, but the RIP router happens to be malfunctioning at that time, without this router update mechanism, other routers will not know that it is currently malfunctioning and still transmit the packet according to the original routing path. The result, of course, is that the packet always cannot reach the destination host. Although it may have been tried many times.

Parsing example of 2.RIP routing periodic update mechanism

In order to better understand the update mechanism of the routing table of the RIP protocol, let's take the simple interconnection network shown in figure 8-6 as an example to discuss how the routing tables in each router in the figure are established.

Figure 8-6 example of RIP routing table establishing a network

(1) at the beginning, the routing tables in all routers have only the routing table entry information for the networks to which they are directly connected. But it is not a RIP routing table entry, it is a directly connected routing table entry, there is no need for the next hop (represented by "- -"), and the metric "distance" is also 0. The initial routing table of each router is shown in figure 8-7, with only two directly connected network routing entries.

(2) next, each router will send routing updates to neighboring routers according to the set period (default is 30 seconds). Which router will send routing updates first depends on which router is turned on first. Assume that router R2 first receives routing updates from routers R1 and R3, and then updates its routing table, as shown in figure 8-8. It can be seen that it adds a new route table entry to the 10.0.0.0 network and 30.0.0.0 network through R1 and R3, respectively, with a metric of 1 because it has only gone through one hop.

Figure 8-7 initial routing tables for R1, R2, and R3

(3) after R2 updates its routing table, it sends the complete routing table to neighboring routers R1 and R3. Routers R1 and R3 are updated respectively. According to the rules for updating the RIP routing table described earlier, R1 first adds 1 to each metric received from R2 if the routing table in figure 8-8 is added to each metric, and the resulting routing table is shown in figure 8-9.

Figure 8-8 R2 routing table after routing update figure 8-9 R1 forms the routing table by adding 1 to the received routing table from R2

(4) then R1 compares the routing table shown in figure 8-9 with its original routing table (shown on the left in figure 8-6). All newly added routing table entries with metrics less than or equal to the original will be updated, and routing table entries with larger metrics will ignore updates. After row comparison, it is found that there are two new routing table entries, whose destination networks are 30.0.0.0 and 40.0.0.0, respectively, which are directly added to the routing table. For the two existing 10.0.0.0 and 20.0.0.0 entries, it is found that the routing metric ("distance") value 1 is greater than the original 0, and the update is ignored, resulting in the updated routing table of R1, as shown in figure 8-10.

Using the same method, you can get the routing table of R3 after receiving routing updates from R2, as shown in figure 8-11. However, there is a problem with RIP routing protocol, that is, the network convergence is relatively slow, when the network failure, it will take a long time to transmit this information to all routers, and there are many invalid routing updates.

Figure 8-10 routing table of R1 after receiving routing updates from R2 figure 8-11 R3 routing table after receiving routing updates from R2

Still taking figure 8-6 as an example, all three routers have established their own stable routing tables, assuming that the connection between the R1 router and Network 1 (10.0.0.0) is disconnected. At this point, R1 can immediately discover and update its routing table, change the distance of the routing table entry to 10.0.0.0 to 16 (that is, unreachable), and send this routing update to R2 30 seconds later. However, the routing update that R2 gets from R3 is "the distance to the 10.0.0.0 network via R2 is 2," and the metric is obviously smaller, so R2 updates this routing table entry to "distance 3 to 10.0.0.0 through R3". It is then sent to R3 through the routing update, and the routing table of R3 is updated to "the distance to the 10.0.0.0 network via R2 is 4". R3 then sends the message to R2 through the routing update, and the result is that "the distance to the 10.0.0.0 network through R3 is 5", which is repeated until the distance of the routing table entry reaches 16. R2 and R3 do not know that the 10.0.0.0 network is unreachable.

In order to solve this deficiency, the split horizon technique is produced, that is, updates to the same routing table entry are no longer sent out the interface that receives the routing table entry, which will be described later in this chapter. The convergence mechanism of RIP routing is described in detail below.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.