Surely the instructor can't be wrong?

I'm struggling to understand the result of an exercise I just saw related to route summarization. I think I understand the concept, but the result leaves me with a question. No matter how frequently or loudly I ask the video my question it never seems to get answered, so I thought I'd try it here.

The scenario is that you have two routers, (A &

connected to a core router. Router A has 172.16.129.0/24, 172.16.130.0/24 & 172.16.131.0/24. Router B has 172.16.193.0/24, 172.16.194.0/24 & 172.16.195.0/24. The question is how best to summarize these routes to the core. The method has you determining how many bits are in common vs how many vary between the networks being routed... I think I get all that. For these networks, that would have the summaries set with 22 bit masks.

The part confusing me is that the result proposed is to supply summaries of 172.16.128.0/22 and 172.16.192.0/22. I can see some logic in this since 129 and 193 are superfluous within 22 bits, but it kinda smells like an overlap to me.

Doesn't this mean that a packet addressed to 172.16.128.55/24 (for instance) could be sent to router A even though there is no network connected to it containing that address? I'm thinking I'd need to account for the fact that 172.16.128.0/24 may well have been assigned elsewhere.

If I'm missing something basic, feel free to just point me. I doubt I'll dwell on it for too long.

Thanks in advance.

Kevin

Find more posts tagged with

Comments

networker050184

Summarization is not an exact science so you sometimes may end up summarizing more than you need to. Just like with IP addressing if you need 11 host addresses you will have to go with a /28 even though it gives you 14 addresses it is the closest fit. Its all about the closest fit!

When summarizing you also must take into account how the router will choose the path. If a router sees routes 172.16.128.0/24 and 172.16.128.0/22 it will choose the longest match (the /24 in this case) so the summarized route, even though it contains that address space, will not be used. It is something you have to be careful with in the real world to keep yourself from some head aches. Always ensure you have more specific routes where needed when summarizing.

billscott92787

It all comes back to sweat sexy binary. Did I say sexy? Yes, LOL.

Below is the break down of binary for each individual route as well as the summary of the routes. I'd go back and review this. Definitely review the circumstances when summarization is best for.

Router A

172.16.129.0 /24
172.16.130.0 /24
172.16.131.0 /24

1. 10101100.00010000.10000001.00000000
2. 10101100.00010000.10000010.00000000
3. 10101100.00010000.10000011.00000000

Summary Route: 10101100.00010000.100000 00.00000000 /22

172.16.128.0 /22

Router B

172.16.193.0 /24
172.16.194.0 /24
172.16.195.0 /24

1. 10101100.00010000.11000001.00000000
2. 10101100.00010000.11000010.00000000
3. 10101100.00010000.11000011.00000000

Summary Route: 10101100.00010000.110000 00.00000000 /22

172.16.192.0 /22

EdTheLad

Yes, you are correct but this question clearly mentions to pick the best summary address for the range of addresses given.The best summary address is a /22 and if you tried to use 172.16.129.0/22 the cisco ios will default you back to 172.16.128.0/22 as 128 is the starting address in this block.In a real world scenario if you were worried about this you could put a static route to the null interface i.e. "ip route 172.16.128.0 255.255.255.0 null 0" , this will discard any pkts received destined for 172.16.128.x
Anyway this will highlight the importance of good address planning.

KPWright

Good Stuff!! And quick.

Thanks for the nearly instant feedback. I'm glad to see my understanding of the subject is ok, even if my knowledge of the implications is still a bit limited.

I didn't know Cisco routers limited summary addressing to blocks.

I had read about longest path, but hadn't related that to this scenario.

I didn't know that you could throw packets to the bit bucket with a route to a null interface.

So I got a much more satisfactory answer than the video was ever going to provide

and learned several somethings in the process.

Thanks again. I've got a feeling you may hear from me again. With any luck, I may even be able to contribute an answer or two in the future.

Kevin

billscott92787

Yeah, the example 172.16.128.0 with /22 mask a /22 mask is 255.255.252.0

So you determine the "block size"

256 - 252 = 4. So each subnet "block" is going to be equal to 4 in size. For example,

172.16.0.0 / 22

172.16.4.0 /22

172.16.8.0 /22

and so on then we finally get to 172.16.128.0

So these are in the same subnet

172.16.128.0 - 172.16.131.0 /22 the ip range between these would be

The IP range for this subnet would be:

172.16.128.1 - 172.16.131.254 and the broadcast of this subnet block would be 172.16.131.255