Mismatched ip addresses!

Unanswered Question
Jul 25th, 2009

Hi !

I found this question and answer in CISCO PRESS ICND2 CHAPTER 12 REVIEW QUESTIONS!

Two routers have a serial link between them,with the link configured to use PPP and with EIGRP configured correctly for all the interfaces.The engineer can ping the IP ADDRESS on the other end of the link,but not the IP ADDRESS of the other router's LAN interface.Which of the following is the likely cause of this problem?

a. The CSU/DSU connected to the other router is not switched ON

b. The serial IP ADDRESS on the router at the other end of the link is not in the same subnet

c. CHAP authentication failed.

d. The router on the other end of the link has been configured to use HDLC.


Explanation given: With PPP, two routers can use IP address in different subnets on opposite ends of the links and a ping to the other router's serial IP address works.However,this subnet match causes routing protocol to fail when forming neighbor relationships to exchange routes ,so neither router learns EIGRP routes from the other.

But in another one famous CISCO literature,I have found that

"If you have a point-to-point link but the encapsulations aren't the same, the link will never come up.

After that it is mentioned in the TOPIC MISMATCHED IP ADDRESSES,

Mismatched IP Addresses

A tricky problem to spot is if you have HDLC or PPP configured on your serial interface but your IP addresses are wrong. Things seem to be just fine because the interfaces will show that they are up. But it won't work since ip addresses are in different subnets."

I found contrary of these 2 answers.

Can anybody please explain?


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
jbrenesj Mon, 07/27/2009 - 12:32

The answer in the ICND2 is fine.

With PPP, each end of the link could be in different subnet and you will be able to ping the other side bot nothing beyond that as the routing protocols won't create a neighbor relationship

so this is hardly used in a real network.

the other answer is also correct but they are saying that "Things seem to be just fine because the interfaces will show that they are up. But it won't work since ip addresses are in different subnets."

--> What won't work here is routing for instance.

hope this helps

KnowledgeAcquirer Tue, 07/28/2009 - 17:11


The question you have above is #7 on p. 435 of the ICND2 book.

Your "Explanation given:" is on pgs. 451-452 and Example 12-7 and is summarized in Table 12-5 on p. 452. The answer B is on p. 640. Please focus on the following:

1) PPP and EIGRP are configured correctly for the 2 routers with a serial link between them. This implies that the serial links are not the problem and further corroborated by the statement that "the engineer can ping the IP ADDRESS on the other end of the link, but not the IP ADDRESS of the other router's LAN interface.

2) Even if the serial links on the 2 routers were in different subnets, with PPP configured on both routers, the pings to the serial interfaces on the 2 routers work because PPP advertises the host routes of the 2 routers.

3) PPP does NOT advertise the host route of the LAN interfaces that reside on the routers. The LAN interface of the other router must reside in a different subnet than the subnet for the serial link between the two routers. The purpose of a router (or any Layer 3 device) is to forward packets between DIFFERENT GROUPS of addresses (i.e, network or subnet). If the LAN interface of the router did not reside in a different subnet, then the use of the router is frivolous.

If you are still confused, please respond and let me know if you have Boson's Network simulator and Network Designer. I'll try and attach a topology and multi-device configuration that demonstrates this.

Yes, at first I was confused when I first read this question while reviewing for my CCNA exam, until I sketched it out and outlined exactly what must be occuring. I also outlined what must/must not occur for each of the answers. Unfortunately, Cisco exams have a short timeframe so I have to get faster at doing this.

shanparames Fri, 07/31/2009 - 06:28

Thankyou very much for your reply!

I understood your reply but can u please send me the topology explaing this concept.I have Boson's Network simulator and Network designer.

And go through this post in "http://www.lammle.com/discussion/showthread.php?p=5858"


Once again thanku very much for your reply,



shanparames Tue, 08/04/2009 - 09:24

Hi Knowledgeacquirer!

Please help me in this regard of boson and the solution you outlined!

With thanks!


KnowledgeAcquirer Tue, 08/04/2009 - 13:42

Hi S. Swaminathan :-)

I tried going to the thread but I can't post there because I don't have a valid login.

I have the topologies and multi-device configs along with a Word/PDF document describing each of the scenarios that I built to test this. I also have a JPG snapshot to show that the Boson N/W simulator does NOT mimic the ppp advertisement behavior as discussed below and does NOT result in the output of Example 12-7 on p. 451 of the ICND2 book.

The keypoint is the 2 sentences on p. 451 (3rd paragraph) of the ICND2 book:

As it turns out, a router using PPP advertises its serial IP interface to the other router, with a /32 prefix, which is a route to reach just that one host. So both routers have a route with which to route packets to the other end of the link, even though two routers on opposite ends of a serial link have mismatched their IP addresses.

If this is true, then the issue is NOT Cisco, but the PPP protocol. Since the router in Example 12-7 is a Layer 3 device with a /24 mask, this doesn't matter since PPP is advertising a /32 mask at Layer 2 (recall that Layer 3 is encapsulated within a Layer 2 frame, in this case a PPP frame).

KnowledgeAcquirer Tue, 08/04/2009 - 14:30

Here is additional info:

If you ever used IPCP for address allocation with PPP (”ip address negotiated” on client side and “peer default ip address” on server side) you may have noticed that the mask assigned to a client is always /32. It does not matter what mask a server uses on it's side of the connection, just PPP is designed to operate this way.

This is from the following link:


But you may want to focus on the display shown on p. 451 for Example 12-7. Specifically, the line that is NOT highlighted: is subnetted, 1 subnets

Notice the /32 mask? That is why the 2nd highlighted line in Example 12-7 shows is directly connected, Serial0/0/1.

Since IPCP (the IP Control Protocol) that is defined as part of PPP defines the /32 mask, it doesn't matter that each of the router's IP addresses use /24 mask from the viewpoint of the serial interface for ppp encapsulation.

HDLC encapsulation does NOT use the ppp encapsulation, instead relying on the IP address assigned to each router (and thus the mask used there, in this case, the /24 mask).


This Discussion