Are these Contiguous Networks?

urvi

Hi Friends
A problem from the Pearson exam(CD accompanied with the Odom book )which i post here-

The above problem shows some subnets . Do these subnets constitute what is called a contiguous network?

Thanks in advance.

101010

I'll take a stab at this and say no, this is not a contiguous network.

I say this because the, if you take a look at this page on Cisco website: Why Doesn't RIP or IGRP Support Discontiguous Networks? - Cisco Systems

The first line in the introductions states:

A discontiguous network comprises a major net that separates another major net.

So the fact that your link between router B and C in ytour example [10.82.113.0 /30] is on a diffarent major network than the rest causes it to be discontiguous. Now as I understand it, if the link had been say 10.25.67.160 /27, it would have remained contiguous.

I am still learning this material myself, so I hope someone with a better understanding might chime in and verify my grasp of the topic.

NetworkVeteran

101010 wrote: »

The first line in the introductions states: A discontiguous network comprises a major net that separates another major net. So the fact that your link between router B and C in ytour example [10.82.113.0 /30] is on a diffarent major network than the rest causes it to be discontiguous.

Good try! Actually, the network is contiguous because the "major" network is 10.0.0.0/8, hence no auto-summarization. From that same link you provided--a good page for sure--Auto-"summarization is done to the major classful boundary."

Why Doesn't RIP or IGRP Support Discontiguous Networks? - Cisco Systems

Answers--

A is correct because it appears in the topology and no auto-summarization would occur.
B/C are excluded because they don't even appear on the topology and hence are silly!
D is excluded because auto-summarization doesn't take place.
E is excluded because C would send rather than receive this in updates.

101010

Thanks for the clarification.

***EDIT***
@NetworkVeteran

So if I am understanding you correctly, in urvi's example if the 10.82.113.0 /30 had been instead a 11.82.113.0 /30 it would then be considered discontiguous as we would have a different major network, due to the change in the /8 section of the network. My mistake was only seeing that the network had changed, but the change had occurred in the "sub-net" section, as opposed to the "major network" section.

Similar to how "Figure 1" in the link I provided. That one is considered discontiguous as the change occured in the "major network" [/16] section.

Feel free to correct me if I am once again wrong.

urvi

Hi 101010,
What is the "/8 section" that you mentioned above?I didn't get you again

101010

I was referring to the fact that in a standard Class A network, the network mask is the first 8 bits or /8.

Standard Classful Masking:

Class A:
/8
255.0.0.0
11111111.0.0.0

Class B:
/16
255.255.0.0
11111111.11111111.0.0

Class C:
/24
255.255.255.0
11111111.11111111.11111111.0

NetworkVeteran

Aye! That is totally correct.

101010

Thank you for the clarification NetworkVeteran. Always awesome when something clicks.

urvi

Thanks NetworkVeteran!
But I have one last question- Does this mean that 10.25.67.64/27 will also be advertised to Router C in the above case?
2ndly,the link you provided implies that RIPv1 don't support discontiguous networks but does RIPv2 support discontiguous networks? Also RIPv1 does support contiguous like RIPv2?
I'm really confused!

FrankGuthrie

Well the 10.25.67.64/27 network will only be correctly advertised to Router C if you turn of auto-summarization for RIPv2. IF not it will be advertised as 10.0.0.0/8.

RIPv1 does NOT support discontinuous networks. This means in the above topology, you would run into problems, because all the networks will be seen as 10.0.0.0/8
RIPv2 does support discontinuous networks, but this is turned on by issuing the #no auto-summary command.

theodoxa

I thought RIPv2 [just like RIPv1] and for that matter every routing protocol other than OSPF auto-summarized unless you explicitly told it not to. Its a huge PITA because auto-summary can break your routing on a well designed network.

For example, when I was studying RIP, I created a Hub and Spoke topology. All the LAN links on the spoke routers were subnetted from a single classful network using VLSM. I quickly discovered that my traffic was not going where it was supposed to because each router was summarizing its links to the clasful boundary. So, the upstream (Hub) router was effectively receiving a route to the same network on all of its links. The traffic for all the subnets was being routed out a single link [which half the time was the wrong link].

iamme4eva

Auto-summarisation only occurs at a classful network boundary.

So in the above topology, all of the networks are a derivative of the classful network 10.0.0.0/8 (class A). No auto-summarisation would occur.

The network becomes discontiguous when you have to cross over another network to get to an address in the same classful network...if there was an 11.0.0.1 in there somewhere or something.

urvi

Hello Frank,
I think you are totally wrong if you think the network I posted above in my first post is a discontiguous network. As correctly pointed by Network Veteran ,it is a contiguous network and please dont confuse me anymore. I hoped NetworkVeteran would reply but he didn't. Instead most of you gave incorrect answers except eva and messed up the situation instead!
p.s.:As far my understanding goes,contiguous networks like the one I posted above can't have auto-summarization.
Only discontiguous networks can have auto summarization.

Correct me if I am wrong.
Thanks

JoshyJ

I believe it to be other way around. Discontiguous networks need to have auto-summarization turned off. Contiguous networks can have auto-summarization turned on.

urvi wrote: »

Hello Frank,
I think you are totally wrong if you think the network I posted above in my first post is a discontiguous network. As correctly pointed by Network Veteran ,it is a contiguous network and please dont confuse me anymore. I hoped NetworkVeteran would reply but he didn't. Instead most of you gave incorrect answers except eva and messed up the situation instead!
p.s.:As far my understanding goes,contiguous networks like the one I posted above can't have auto-summarization.
Only discontiguous networks can have auto summarization.

Correct me if I am wrong.
Thanks

urvi

Ya JoshyJ on that I am wrong maybe. But the network is contiguous

iamme4eva

The network IS contiguous, as there is only ONE classful network. A discontiguous network is as I defined in my previous post.

The key to this is understanding what auto summarisation does, and where it does it.

Auto summarisation will automatically summarise a subnet back to it's classful network, AT A CLASSFUL NETWORK BOUNDARY.

So 10.1.1.0/24 would be summarised back to 10.0.0.0/8...but ONLY at the border of another classful network (so in this example, something that doesn't start with 10.)

If you think about that concept and apply it, then if you have a contiguous network auto summary doesn't matter. If it is contiguous, then there will only be one router advertising the summary route, so all is good. In the network diagram urvi has presented, there will be no auto-summarisation, as there are no classful network boundaries to cross, so it can be off or on.

Auto-summarisation is a bad thing for discontiguous networks, because if two routers advertise the 10.0.0.0/8 summary, then only half your network traffic will get to where you aim! I'd draw a picture, but I'm at work.

urvi

Hi eva,I think what you are trying to say is that all the contiguous subnets should be only on one side of the router for auto-summary to work? am i correct?
Most people on networking sites dont make the above thing explicitly clear which had me confused as well. Now I see how contiguous sites can have auto-summary only if they are all on one side of the router.

iamme4eva

An entire classful subnet should be at one side of a router, which is what makes it contiguous. The word contiguous literally means "alongside or touching".

Good network design involves keeping your similar subnets as close together as you can anyway to make for better summarisation - be that automatic or manual.

Carpe Porcus

This should clear up any misunderstanding:

https://learningnetwork.cisco.com/servlet/JiveServlet/previewBody/20977-102-6-75703/Behavior%20of%20RIP%20version%201%20in%20contiguous%20and%20discontiguous%20networks.pdf.