Answer: Overlapping Connected and Routing Protocol Routes
This post wraps up the #CCNA Q&A focused on how routers add routes to their routing tables. What happens when a router happens to learn three separate routes for the same subnet ID β but with different masks? And how does yet another overlapping subnet β same subnet ID, different mask β affect the routerβs logic, if that route is that special type of route for a connected subnet? Todayβs post walks through the answer to the question and the reasons.
Answer
F
Disclaimer
(Same disclaimer as the previous question!) First, for emphasis, the design used for this question is poor and not recommended. I just used it as a way to make one router receive routing updates for three subnets that have the same subnet ID but different masks, as learned with three different routing protocols, just as an exercise. Donβt use a similar design in a real network! The goal of the question is to let us focus on how a router thinks about adding routes to its IP routing table.
Background
Compared to the previous question and answer, this question just adds a connected route for subnet 10.1.2.0/24, to the other three learned routes. This additional route begs the question: does a connected route somehow change the rules a router uses when choosing what routes to add to the routing table? The short answer: no.
To review, a router must think about how to choose amongst competing routes to the same subnet. The subnet is not considered the βsame subnetβ unless both the subnet ID and mask are the same. In this question, and the previous question, none of the subnets were the same subnet! As a result, the router never had to use any logic to decide which of the multiple routes to the same subnet was better, because the router knew of only one route to each individual subnet.
The Correct Answer
The correct answer, F, states that the router will add routes to all four subnets. None of the subnets β including the connected subnet β is the exact same subnet. So, the router simply adds routes for each.
Butβ¦ Butβ¦
If youβre already thinking this, great! But what would have happened if the question had made all four of those subnets use the same mask? If that were the case:
- All four subnets would appear to be the same subnet: the same subnet ID and same mask.
- The router would have to choose between all four competing routes
- The router would use the administrative distance of the route to choose the best route, so the connected route would win.
Hello, had a quick question for you unrelated to this question on overlapping subnets. The old CCNA exam that expired in September/October of 2013 happened as I was studying to take the exam. It changed and I had to get the new book for the new ICND1 exam. Which I passed in February of 2014. π My question is… Are these new exams going to change in the near future or will it be another 5 years before another change is made. I haven’t been able to find any information regarding this so I’m assuming that no announcement has been made in regards to this. I know Cisco can change the exams whenever they would like to. Do you know where they would post this information first if they plan on changing the exams again? Any insight would be gratefully appreciated. By the way, love the books! π
Hi Jeff,
Thanks! Glad you’re loving the books!
Cisco has not announced changes to the exams compared to the 100-101, 200-101, and 200-120 exams. When they do, the first public information would be somewhere at cisco.com, usually at cisco.com/go/cretifications. You can/should also watch at the Cisco Learning Network, learningnetwork.cisco.com. That said, if you watch CLN, people will ask, and speculate, but Cisco won’t announce new exams until they are ready. When Cisco announces, there is usually 6 months (sometimes longer) before the old exams go away, and a migration path if you’re part way through passing all the exams for a given cert. That’s the best way to watch for when Cisco announces cert changes to any of their certs.
Hope this helps!
Wendell
Hi Wendell,
I’ve loved your books too (thumbs-up) so far. Great work.
On to the question. I really laughed when I read the “But… But” BUT mine is slightly different.
What will happen if a packet arrives at R1 with destination IPv4 address of 10.1.2.1 and why?
I simulated the lab in PT and the packet chooses the “EIGRP-path” if R1’s LAN is not set and “Connected-path” with R1’s LAN configured using /24.
Thank you,
AK
AK,
Wow, this is an oldie but a goodie! Glad it gave you a chuckle.
Short version: packet is forwarded with the most specific route, so it’ll match the /27 route to R4. That is, as listed here, R1 has a route to all four 10.1.2.0 subnets (each with different masks), but when matching the table, it uses the most specific. In particular, the admin distance is not considered at that point; it’s only considered when routes to identical prefix/mask combinations exist, when making the choice of which to put into the table.
W
Hi Wendell,
Thanks for these questions, they really are useful.
Just noticed what must be a typo in the question that changes the answer. If the interface is configured with “ip address 10.1.2.0 255.255.255.0” the command will be rejected as it’s not a valid up address, so it’s connected route is not added.
Probably it was intended to be “ip address 10.1.2.1 255.255.255.0”.
Karim,
You are 100% correct. Fixed it as you suggested. Now it reads as I intended, rather than how I typed it. π Thanks much.