…well, I don’t want to list the answer to yesterday’s exam question right up front, spoiling the chance for anyone who didn’t yet read yesterday’s blog entry (http://www.networkworld.comhttps://www.networkworld.com//community/node/19364) to have a chance to play along and time themselves answering the question. But as of the time I posted today, yesterday’s unscientific survey shows 5 of 20 people with over 1:30 to convince themselves of the answer – which is probably long enough so that you’d be faced with guessing the answer and moving on, or risking getting under more time pressure if you kept plugging away at the question. So, the literal answer is at the bottom of the post – so scroll to the bottom if you like. I’ll step through the strategies for attacking this question for today’s blog.
The question text itself gave some really important clues right off the bat. The question said that there were no layer 1 or layer 2 problems, and that a ping (PC1 to PC3) didn’t work – which should tell you right away that there’s something wrong at layer 3. So, taking a more methodical view of the problem, you can either first look at the IP data plane (aka the IP packet forwarding process) or control plane (aka looking at how the routers learn the routes, in this case with RIP). I’ll look at the data plane first here.
There are 6 required steps for a ping from PC1 to PC3’s IP address to work – three for sending the ICMP Echo from PC1 to PC3, and three for PC3 sending the Echo reply back to PC1. First, to get the packet to PC3:
1) PC1 must be able to send packets to PC1’s default gateway, presumably R1’s Fa0/0 IP address (10.4.102.111)
2) R1 must be able to forward the packet (based on its routing table) to R3’s next-hop address of 10.4.25.33
3) R3 must be able to forward the packet to PC3 (10.4.133.133) over the LAN
Then, for PC3 to send the ICMP Echo Reply back to PC1:
4) PC3 must be able to send packets to PC1’s default gateway, presumably R3’s Fa0/0 IP address (10.4.135.33)
5) R3 must be able to forward the packet (based on its routing table) to R1’s next-hop address of 10.4.20.1
6) R1 must be able to forward the packet to PC1 (10.4.98.1) over the LAN
Also, you need to keep in mind some basic IP addressing rules when examining the problem – mainly, that hosts on the same LAN/VLAN should be in the same subnet, that routers on the same serial link should be in the same subnet, but that IP addresses on different LANs and serial links should be in different subnets.
So, everything I’ve mentioned so far should only take a few seconds to realize once you’re studied up for the exam, because all these are foundational concepts related to IP and routing. Now let’s take a look at this specific problem. Thinking about the 6 steps in order:
Step 1: PC1 should be able to send packets to its default gateway, because if it couldn’t, PC1’s ping of PC2 would also have failed, and the question said a ping from PC1 to PC2 worked. Also, some math would confirm that PC1’s IP address and R1’s Fa0/0 IP address are both in subnet 10.4.96.0 255.255.248.0.
Step 2: The serial link from R1 to R3 shows IP addresses of 10.4.20.1 255.255.248.0 and 10.4.25.33 255.255.248.0, respectively. After a bit of math, you can see that R1’s IP address is 10.4.16.0 255.255.248.0, and R3’s is in subnet 10.4.24.0 255.255.248.0. (For convenience, you could have changed the mask to use prefix notation, or /21.) Anyway, R1 would not be able to forward the packet to R3 because of these mismatched IP addresses – so that’s the first possible root cause of the problem.
Step 3: to make sure no other problems exist, you could check R3’s LAN IP address and PC3’s IP address, and find that they both reside in subnet 10.4.128.0/21.
Steps 4-6: The IP addresses have already been examined, so you don’t need to look for IP subnet mismatches with the IP addresses. For these steps, you could check out the IP routing tables on the routers (if they were supplied; I didn’t put them in the blog yesterday.)
You could go on to examine the control plane, looking for routing protocol issues or problems with any static routes. One hint with this question was that the question didn’t supply the routers’ routing tables, plus the RIP configurations on the three routers are very simple, and complete. So, if you were being thorough, you could have spent some time looking at the RIP config, confirm that it matched all interfaces shown in the figure (which it does), and rule out RIP as being a cause of the problem.
So, if this question was on the real exam, and they asked for a single answer (the exam tells you how many correct answers each question has), and you got to the realization of the IP addressing issue on the R1-R3 serial link, then you’re done – other than checking the other answers a little more deeply to make sure you didn’t miss something.
So, back to the excuse to look at this question – how much time does it take? I’d be interested in some opinions about a few things regarding the question. Did any of you think about the RIP config in detail first, before thinking about the routing/forwarding path? Did any of you find the R1-R3 IP address mismatch, but then kept the timer running while you made sure that was the problem – and if so, how much longer? Anyone have any other suggestions about how you’d attack this particular problem? And, other than just taking time to calculate the subnet number for the various IP addresses, was there anything else that seemed to take too much time?
Tomorrow, I’ll close the week with some discussion of some general strategies for dealing with the time pressure on the exams.
PS ANSWER: “R1 and R3 are configured to be in different subnets, but they should be configured to be in the same subnet”
