Classful routing protocol behavior

kpjungle

Hi,

Trying to dig a bit deeper into classful routing routing protocol behavior, and read this snippet in BSCI self-study:

"A router sends the entire subnet address when an update packet involves a subnet of the same classful network as the IP address of the transmitting interface. The Receiving router then assumes that the subnet in the update and the interface use the same subnet mask."

The lab's ive done to test this out shows that if a subnet of the same classful network (ie. subnet 172.16.2.0 has classful net 172.16.0.0), and has the same subnetmask as the receiving interface it gets entered into the routing table. If it is of a different subnetmask, it wont get into the routing table because the router summarizes it on the class (172.16.0.0). So what I take issue with here, is the last part about the receiving router. It will enter it a subnet of the same classful net into the routing table, as long as the subnetmask is the same.

Am I reading it incorrectly, or does the information not match up?

tech-airman

kpjungle wrote:

Hi,

Trying to dig a bit deeper into classful routing routing protocol behavior, and read this snippet in BSCI self-study:

"A router sends the entire subnet address when an update packet involves a subnet of the same classful network as the IP address of the transmitting interface. The Receiving router then assumes that the subnet in the update and the interface use the same subnet mask."

The lab's ive done to test this out shows that if a subnet of the same classful network (ie. subnet 172.16.2.0 has classful net 172.16.0.0), and has the same subnetmask as the receiving interface it gets entered into the routing table. If it is of a different subnetmask, it wont get into the routing table because the router summarizes it on the class (172.16.0.0). So what I take issue with here, is the last part about the receiving router. It will enter it a subnet of the same classful net into the routing table, as long as the subnetmask is the same.

Am I reading it incorrectly, or does the information not match up?

kpjungle,

Do you have "(config)#ip classless" configured on the receiving router?

dtlokee

The mask is not what is important here, it's the classful network of the interface. If you have an interface advertising 172.16.2.0/24 out an interface that has an IP address of 172.16.3.1/30 it will advertise the route as 172.16.2.0/24 because the update is sent out an interface with the same classful network. Now if you sent an update for 172.16.2.0/24 out an interface with an IP address of 172.17.1.1/24 it would advertise the route as 172.16.0.0/16 to the neighbor.

singh8281

dtlokee wrote:

Now if you sent an update for 172.16.2.0/24 out an interface with an IP address of 172.17.1.1/24 it would advertise the route as 172.16.0.0/16 to the neighbor.

In case of EIGRP, Unless auto summary is enabled the route 172.16.2.0 will be installed in neighbor routers.

kpjungle

dtlokee wrote:

The mask is not what is important here, it's the classful network of the interface. If you have an interface advertising 172.16.2.0/24 out an interface that has an IP address of 172.16.3.1/30 it will advertise the route as 172.16.2.0/24 because the update is sent out an interface with the same classful network. Now if you sent an update for 172.16.2.0/24 out an interface with an IP address of 172.17.1.1/24 it would advertise the route as 172.16.0.0/16 to the neighbor.

I cant see this happening in my lab.

If i advertise 172.16.2.0/24 over an interface with ip 172.16.1.1/30, it wont send it to the other router.

R1 = Sending router.
R2 = Receiving router.

output from R1:
00:11:47: RIP: sending v1 update to 255.255.255.255 via Serial0/0 (172.16.1.1)
00:11:47: RIP: build update entries - suppressing null update
00:11:47: RIP: sending v1 update to 255.255.255.255 via Loopback0 (172.16.2.1)
00:11:47: RIP: build update entries - suppressing null update

However, if I change the advertised route, to this mask: 172.16.2.1/30 (same mask as the outgoing interface), it works (it actually sends the route):

output from R1:
00:14:27: RIP: sending v1 update to 255.255.255.255 via Serial0/0 (172.16.1.1)
00:14:27: RIP: build update entries
00:14:27: subnet 172.16.2.0 metric 1

output from R2:
00:15:25: RIP: received v1 update from 172.16.1.1 on Serial0/0
00:15:25: 172.16.2.0 in 1 hops

and it then appears in the routing table on R2:
172.16.0.0/30 is subnetted, 2 subnets
C 172.16.1.0 is directly connected, Serial0/0
R 172.16.2.0 [120/1] via 172.16.1.1, 00:00:13, Serial0/0


So to sum up my experience in the lab it would point to:
1) if the advertised route is of the same major net (172.16.0.0/16), and has the same subnet mask as the advertising interface, it will advertise it, and the receiving router will enter it into the routing table.
2) if the advertised route is of the same major net (172.16.0.0/16), but has a different subnet mask than the advertising interface, it will not send it (supressing the update).
3) any other major net (be 172.17.0.0/16 or 10.0.0.0/, it will send the major net and the receiving router will install it with the major net only.

In case of EIGRP, Unless auto summary is enabled the route 172.16.2.0 will be installed in neighbor routers.

yes, if it was not turned off it wont be installed, because 172.17 and 172.16 are two different major nets, it will send 172.16.0.0/16.
Also, and interesting bit. EIGRP, which is a classles routing, behaves differently than rip. Even if the subnet being advertised is a subnet of the same major net as the advertising interface, but with a different subnet, no matter summary/no autosummary, it will still insert the routes correctly on the other router.


Thats what the magic lab is telling me

APA

if it is part of the same classful network there is no need to summarize at the major network boundary hence why it sends the update...

gojericho0

Take a look at this link, it may help to explain it a little better:

http://www.cisco.com/en/US/tech/tk364/technologies_tech_note09186a0080093fd8.shtml

gojericho0

Also, as far as I can tell auto-summary on classess routing protocols will have the same effect, unless you are supernetting.