RIPv2 Question

drew2000

Hi All,
The following practice question is driving me nuts!

Imagine an internetwork that has a dozen interconnected routers. The IP addresses of all interfaces on all routers begin with 10.1. RIP version 2 has been chosen as the routing protocol. Which three commands must be configured on all routers to enable RIP V2?

A. "network 10.1.0.0"
B. "no version 1"
C. "network 10.0.0.0"
D. "router rip version 2"
E. "version 2"
F. " router rip"

Correct answer: C, E, F (my answer was A, E, F)

I thought RIPv2 was classless. If so, why would you use the command "network 10.0.0.0" to add a classful network to the routing table, when you could use "network 10.1.0.0"?

Thanks,
Drew

kalebksp

In RIP, whether version 1 or 2, the network command is always specified as the classful network. This is actually a particular annoyance when you have two networks, say 10.1.2.0/24 and 10.1.5.0/24, and only want to advertise 10.1.2.0/24. You must use "network 10.0.0.0" to advertise 10.1.2.0/24, in which case you can't stop RIP from advertising 10.1.5.0/24 as well.

Of course there is a way around advertising 10.1.5.0, but it's CCNP level stuff.

phoeneous

I got into a heavy debate with someone over this, mainly because Im hard headed. When it comes to rip, just remember that the network command is always classful. It doesnt actually advertise the route like most people think, it allows the interfaces that fall under the network command scope to take part in the rip process. When you run the sh ip route command, you will see classless routes listed.

drew2000

Ok cool, I actually like these explanations because it's easy (nothing crazy I don't understand).

Thanks guys,

Andrew

daniel2009

RIPv1 is classful
Back in the old days, networks were made with classes, one can not vary the network size except between 256, 65536 or 16 million hosts, subnet mask is not sent in routing updates.
Thus the whole network will be on the same subnet-mask.

Class A has subnetmask 255.0.0.0,
Class B has subnetmask 255.255.0.0
Class C has 255.255.255.0

Notice here the amount of 0's in the subnetmask. In class C the last byte (8 bits) is 0. 2^8 = 256. Thus this one has 256 hosts
In Class B there are 2 bytes which are 0, 2 * 8 bits = 16 bits. 2^16 = 65536 hosts.

The subnet are also written as /8, /16 and /24.
This is called the network prefix, the numbers are the amount of bits used.
For example:
class A 255.0.0.0 = (binary) 11111111.00000000.0000000.00000000
only 8 bits are used, so /8.

So there can be only 3 types of network sizes, the subnet mask length is not variable. No VLSM (variable length subnet mask) support

I've no idea why classful routing protocols are written about in ccna, as they are out-dated. I mean, for example, why would one create a network with 16 million free IPs when only 7000 are needed?

RIPv2 is classless.
one can vary the network size using different subnetmask length
RIPv2 has VLSM (variable length subnet mask), so the length of the subnet mask can be any amount of bits, for example /9 or /11. (11 "1" bits from left)
This means you can create networks of a preferred size and not waste to many IPs.
The subnetmask is sent in routing updates.


Both RIPv1 and RIPv2 routing protocols use hop count as routing metric.
If you are familiar with the programming language C, it might be interesting to look at the program "routed". This is an open-source program that implements rip version 1. it are about 10 code files, all the things like hop count and such are defined in a header file, i believe it was routed.h

rwwest7

Also keep in mind that RIPv2 is classfull by default until you disable auto-summary.

drew2000

Thanks for all of the tips!

Andrew

phoeneous

daniel2009 wrote: »

RIPv2 is classless.

But the network command in either version is still classful. The only difference is that in v2 you will see classless routes in the routing table.

drew2000

I can't thank you guys enough for explaining this to me. I really understand it now.

Drew