rip and subnets

JSW77

i have a question about rip finding subnets. in todd lammles book, he gives an example of several networks connected by routers running rip. these networks are 172.16.20, 172.16.30.0 etc...He points out that you only need to enter network 172.16.0.0 into each router (after "router rip") and the router will "find" the subnets. (and after entering "show ip route" the subnets will be shown) Now i know that a router running rip is aware of directly connected subnets, but how does it find out about the subnets a hop or two away? doesnt a router running rip summarize the subnets as 172.16.0.0(the classful boundry) before sending the update to the next router?

henrock2

Because in the routing table updates, it list the ports, ip address, and other information. The IOS takes that information and puts it all in the update tables. You can put the subnet number in and the IOS changes it to the network number. During the process it test matches, on one or more of the interfaces. The router broadcast or multicast route updates out an interface, listens for incoming updates on the same interface, when sending an update includes the subnet off that infeteface in the routing update.

So basically the router ask" which one of my interfaces matches this network number?" then the protocol uses the metric, in this case, rip uses hop count, which is default to 15, it knows automatcially where to go by the updates. Its built in, unless otherwise specified, to just go, until it reaches 16 hops then it fails, and thus you have a routing loop, then it does whatever it needs to converge.

I hope that helped.

JSW77

yes, i understand that this is how network information is normally propagated. but rip does not send subnet mask info in the routing update (classfull), therefore it cannot specify that the network portion of the routing update should be 172.16.30.0 instead of 172.16.0.0. So it seems to me that when a router sends an update to another router including the networks 172.16.20.0 and 172.16.30.0, they would be summarized classfully as 172.16.0.0 and therefore the next router would not know of the two separate subnets. however in todd lammles book, he has an example where several routers are set up like this:
A

---

B

---

C

---

D
Each network between routers is 172.16.10.0, 172.16.20.0, etc...Now when configuring rip, he states that the only network that needs to be specified is 172.16.0.0. Then he says the router will just "find out" about the other networks. My question is, how does the router just "find out" about the other networks considering that they are SUBnetworks and rip is a CLASSFULL protocol?

henrock2

rip v1 doesnt send subnet mask, but RiPv2 does. Once again, when you configure your interfaces, they are on different subnets, unless your dealing with CIDR. So when the routing protocol gathers the information it gathers the ports, hence using the ANDing process seperate node numbers from subnet numbers. Once that is done, it is put in the routing table and this is how the router knows, which hop to take and where to go. This is how, because the router knows that when you use rip you put directly connected network,how do you come up with the directly connected networks? because they are the interfaces connected "ON" your router, and like i said if you are using classful addressing you put different node addresses for each interface.

JSW77

bump.