Need help in how to approach a CCNA exam question

Unanswered Question
May 7th, 2009

Hi, all: I'm working toward a CCNA, and I'm struggling with a specific type of question. Essentially, it's the: 'You have four sites: site 1 with 500 hosts, site 2 with 100 hosts, site 3 ...' The question asks that you develop a subnet scheme for the sites, and gives all of the correct answers plus a few incorrect answers to choose from. So here's the problem... I can set up a subnet scheme with VLSM all day - that's not too difficult. The number of possible permutations of subnets, however is huge. Trying to wade through all of the different ways to sort and reorder subnets in a VLSM table to figure out one specific answer is just too time consuming. I can't help but think I'm missing the correct approach to solving this type of problem. Any suggestions? Thanks.


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)


I believe they are asking you to configure network addressing to maximize your IP space. I'll clarify:

When they ask:

a site with 500 hosts

They mean:

Configure the smallest network addressing scheme that will allow for 500 hosts.

So you wouldn't want to use:

They want: which allows for 510 hosts.

It is the SMALLEST subnet to meet the requirements.

richlohman Thu, 05/07/2009 - 09:01

Many thanks for the speedy reply. I get the idea of applying masks for maximum efficiency - no worries there. Perhaps I did not explain clearly where I'm having trouble. I fully understand the theory behind what's going on here. The problem is in how the question is designed. For example, if someone said, here's an IP and mask you have to work with - and here is a list of sites with a given number of users for each site - now go to your VLSM table & subnet this out. I can do that. But when I do that, there is more than one possible answer. In fact, there may be hundreds or thousands of possible ways to subnet that out (given the number of sites, users at each site, and the size of the original IP space one has to work with). The problem I'm having is in how the question is designed - they provide *one* answer, and the test taker is required to figure out how to arrive at *that one* answer. I worked out a way to do it, but it takes a long time, and doesn't scale well. That has me thinking that there must be a better approach to solving the problem.

richlohman Thu, 05/07/2009 - 12:16

Sure. For the image, see This image lists 4 sites: New York (20 hosts), Chicago (15 hosts), Dallas (55 hosts), and Seattle (41 hosts). (NY contains the link to the internet, all other sites hang off NY). We are given the block to work with. We need to subnet the block out to fulfill the host requirements at each location. Below the image, we are given 7 possible subnets, of which only 4 are correct. We are to put the correct subnets in the correct location. In every question like this I have seen, one is typically a given, as it's the only one with a mask that matches a specific site. So, here are our options:








Not too difficult. B goes to Dallas, E goes to Seattle, D goes to New York, and G goes to Chicago. A, C, and F are discarded.

So here's a question that gets to the problem of solving this type of exam question: How many combinations exist, which satisfy the question? We could do any of the following, and still have a completely efficient VLSM structure:





- or -





- or -





and on and on...

So, with that, trying to manually construct a VLSM table for every possible combination to meet the criteria set forth in the question, and then select the table that matches the possible answers provided is impossible - you'd never finish the exam. What I'm looking for is the fastest method or algorithm, if you will, to solving the problem presented on the exam. In other words, when you see this question, what's the first thing that has to be done to solve the problem. What's the second, and so on.

So there you have it. For me this is not a subnetting problem, or a VLSM problem, but rather a - problem problem.

jason.chilton Thu, 05/07/2009 - 14:09

I understand where you're coming from here. There really are many ways to subnet the same space and end up with the same amount of used/unused address spaces within.

From all my experiences on these types of questions, unless otherwise specified, it's best to peel off your largest subnet first -- in this case your two /26 blocks, starting at the beginning of your given range. Then work your way on down to smaller subnets.

I think if you follow that method you'll always find the subnets you come up with are the correct answers.

richlohman Thu, 05/07/2009 - 17:28

Whew - I was beginning to think I was losing it. I had the idea of starting with the largest and going to the smallest as it eventually *should* leave the largest contiguous block when you're done; I just didn't know if that was the best approach on these questions. I'll give it a try. Thanks.



This Discussion