Saturday, April 3, 2010


Routing protocols are like gossip. It's like you telling your friends what you know. In this case, what you know are the subnets that you are connected to. It allows routers to automatically builds paths around the network.
There are two (or three) types of routing protocols:
Distance Vector
-Easy to configure
-Not many features

Link State
-Difficult to configure
-Many features

-A mix of both distance vector and link state characteristics

Examples of Distance Vector routing protocols are RIP and IGRP. Link State routing protocols are OSPF and IS-IS. Link State protocols also have higher maintenance. Hybrid protocols like EIGRP are also known as Advanced Distance Vector protocols. EIGRP is proprietary.

Distance Vector protocols are typically slower to converge compared to Link State protocols. Where most Link State protocols are driven by triggered-updates, Distance Vector protocols update periodically.

RIP's algorithm was first developed in 1969. It is called the Bellman-Ford Algorithm. RIP comes in two versions, RIPv1 and RIPv2.

RIPv1 is classful (it advertises networks without advertising the mask). A network would be advertised as without the mask which turns it into Unless a router has another 172.16.x.0/24 network, it would assume that the network has a mask of You can do subnetting, but you must stick to the same mask throughout the network. This means that you can't do VLSM. RIPv1 also does not support authentication. This allows people to easily poison the routing table. Updates are also done through broadcasts which are bandwidth-hungry.

RIPv2 is classless. It supports VLSM. Subnet information is actually sent. It also supports authentication which uses a pre-shared key. RIPv2 also uses multicast address This will not bother any other non-RIP devices with the updates.

We're going to use this topology right now:

We'll start with RA first. To turn on RIP, go to global configuration and type:
router rip

Now you can change the version:
ver 2

Finally, you enter the network statements:

The network command does two things:
1) Tells RIP what networks and interfaces to advertise
2) Tells RIP what interfaces to listen to

Now do the same thing on RB:

Now when you do a "show ip route" on RB, you'll see this: [120/1]

The 120 stands for the Administrative Distance, and the number beside it stands for the metric which is the number of hops away. If somehow the route is learned through a better routing protocol like OSPF, it would have an Administrative Distance of 110. The lower the AD, the more believable is the route. Therefore, if OSPF learns the route, it would replace the route learned through RIP. The number of hops are the number of routers the packet goes through.

You can actually pipe outputs into filters. For example, if you want to only show "ip route" commands in a "show run" output, use:
show run | inc ip route

The Administrative Distance of a static route is actually 1, and connected routes are 0. Therefore, if you have a static route and an advertised route, the static route will take precedence. Do remember that RIP networks are to be entered classfully (if you actually type the non-classful address, it would fix it for us. This however does NOT happen in examination simulations). For individual subnets to be shown, disable auto-summary through:
no auto

This is to be done on all routers running RIP. To see the routing protocols running on a router, use:
show ip prot

To see real-time information about RIP, use:
debug ip rip

To stop all debugs, use:
u all

No comments :

Post a Comment