Learning Subnetting/VLSM

techtips

Gawd
Joined
Jan 3, 2011
Messages
530
Well guys, I'm ashamed to say it but I am in my final CCNA WAN class, I can write my CCNA at the end of July and I've suffered with Subnetting and VLSM since the beginning

I'm doing my CCNA case study and I've just realized how confused I am at it.
Does anybody have any good sites/links that explain each step by step in plain easy to learn English?

I have the IP Addressing and subnetting and VLSM workbooks, which combined had about 140-150 pages, which I do but I find that the techniques used in the workbooks are very drawn out and very long.
Any help would be greatly appreciateda
 
There are several sites I just searched up by googling. I'm sure you can find one that explains it to your liking. The problem is there is no plain English involved. It's all math.

When I started, I viewed every subnet in relation to a class-C, or /24, or 255.255.255.0 mask. Same thing. Then I picture moving the dotted line one way or another along the network/host boundary.

If I need less than 256 addresses (minus network and broadcast), I can cut it in half. That gives me 128 addresses. Make it a /25, or call the mask 255.255.255.128. Cut it in half again, and I have a /26, or 64 addresses, or a 255.255.255.192 mask. (Half of 128 is 64, so adding 64 to 128 gives me 192). Now my original class-C network is in quarters.

eg. 192.168.1.0, 192.168.1.64, 192.168.1.128, 192.168.1.192

You can keep going...make it 8 networks, or 16, etc. As you double the networks, the hosts get cut in half.
Find a cheat sheet and try to reverse follow the logic.


Now, what if you want bigger than a class-C? Go the other direction. Add two class-C networks together and call it 512 addresses, or a /23, or 255.255.254.0. Double that again for four class-C networks combined. That's 1024 addresses, or a /22 or a 255.255.252.0 mask. See how the mask works? Subtracting 1, 2, 4, 8, 16, etc.

Doing the math that way always helped me. It might be convoluted for others, but I'm throwing it out there because you never know what method will click in someone's brain.

Yes, this doesn't really explain the theory of bits, zero's and one's, which you probably learned. It's good to know that, but I don't use it in my daily practice.

If any of this confuses you, forget I wrote it.
 
I found doing it by powers of 2 to work well for me.

How many hosts are on a /18 network? 32 - 18 = 14. 2^14=16, 384 - 2 for the subnet id and mask = 16,382

You need 50 hosts per subnet. What subnet do you use? 2^6 = 64 which is the closest without going under that can hold 50 hosts. 64 hosts = /26 or 255.255.255.192
 
Last edited:
What are some ways that you have tried to learn it? I find people end up finding one style that works best for them. While some of my classmates were learning it by "moving the fence" I found it easiest if I wrote out the binary equivalent of everything.

If you can say what's not working for you maybe we can suggest an alternative.
 
There are several sites I just searched up by googling. I'm sure you can find one that explains it to your liking. The problem is there is no plain English involved. It's all math.

When I started, I viewed every subnet in relation to a class-C, or /24, or 255.255.255.0 mask. Same thing. Then I picture moving the dotted line one way or another along the network/host boundary.

If I need less than 256 addresses (minus network and broadcast), I can cut it in half. That gives me 128 addresses. Make it a /25, or call the mask 255.255.255.128. Cut it in half again, and I have a /26, or 64 addresses, or a 255.255.255.192 mask. (Half of 128 is 64, so adding 64 to 128 gives me 192). Now my original class-C network is in quarters.

eg. 192.168.1.0, 192.168.1.64, 192.168.1.128, 192.168.1.192

You can keep going...make it 8 networks, or 16, etc. As you double the networks, the hosts get cut in half.
Find a cheat sheet and try to reverse follow the logic.


Now, what if you want bigger than a class-C? Go the other direction. Add two class-C networks together and call it 512 addresses, or a /23, or 255.255.254.0. Double that again for four class-C networks combined. That's 1024 addresses, or a /22 or a 255.255.252.0 mask. See how the mask works? Subtracting 1, 2, 4, 8, 16, etc.

Doing the math that way always helped me. It might be convoluted for others, but I'm throwing it out there because you never know what method will click in someone's brain.

Yes, this doesn't really explain the theory of bits, zero's and one's, which you probably learned. It's good to know that, but I don't use it in my daily practice.

If any of this confuses you, forget I wrote it.

Interesting... I've never heard it explained like that. I like that. I think that might be easier to ease people into subnetting by explaining it like that and THEN going into the guts of bit borrowing and subtraction.
 
I found doing it by powers of 2 to work well for me.

How many hosts are on a /18 network? 32 - 18 = 14. 2^14=16, 384 - 2 for the subnet id and mask = 16,382

You need 50 hosts per subnet. What subnet do you use? 2^6 = 64 which is the closest without going under that can hold 50 hosts. 64 hosts = /26 or 255.255.255.192

Call me a calculator dependent idiot, but whats the quickest way in figuring the powers of 2 in your head or on paper?

I'm tired of reading & trying to learn all of these intricate subnetting techniques, the method you described has worked best for me (more in-depth here: http://www.techexams.net/forums/ccna-ccent/38772-subnetting-made-easy.html), so I'd like to make sure I continue to practice with it :)

Edit: Here's a pretty good table I just stumbled upon. I suppose memorizing it up to 2^24 would be sufficient
http://www.vaughns-1-pagers.com/computer/powers-of-2.htm
 
It's basically just doubling the number. I usually do up to 65536 (2^16) in my head. I also write them down as well. 2 4 8 16 32 64 128 256 1024 2048 4096 8192 16384 32768 65536. Just think of RAM :d.
 
It's basically just doubling the number. I usually do up to 65536 (2^16) in my head. I also write them down as well. 2 4 8 16 32 64 128 256 1024 2048 4096 8192 16384 32768 65536. Just think of RAM :d.

512 feels left out. :p
 
Hey everybody, so I've been working on subnetting and VLSM since posting this thread and using a friend's method I feel confident with it.

I am working with some group members and planning out some addresses for a few sites and VLANs.

Here is what I have come up with:
e6de3cec.png


Here is what the rest of my group has come up with:
4bfcf779.png



I have verified with a few other people and mine looks correct....without any issues that I nor they can spot.

My group members work, it seems a bit off with the network addresses lol and some of the address ranges


Mind giving some feedback please?
 
The second one, the subnet ID is off for all of the networks. For example, they give 10.10.11.0 /30 so that gives you a range of 10.10.11.1 - 10.10.11.2. They say the router IP is 10.10.11.238, but that's on a different subnet.

Apart from that, yours looks OK.
 
I just realized I had an error with the Single Large Subnet address on my page, of 10.10.8.0 /22... which should be /23.
 
Interesting... I've never heard it explained like that. I like that. I think that might be easier to ease people into subnetting by explaining it like that and THEN going into the guts of bit borrowing and subtraction.

Thanks. It helps that I only have to memorize the classfull boundaries, then go from there. So I remember /8, /16 and /24. Everything else is math up or down from there.

If I know that a /16 is 255.255.0.0, then it becomes easier to figure out /17 is 255.255.128.0, /18 is 255.255.192.0, etc. I usually work out the hosts and networks in my head, in an equally convoluted way. Everything is a power of 2, or half.

Daily, I usually deal with /21 through /30, so I ended up just memorizing all those based on years of use. I think most of us do...but you have to start somewhere.

techtips, It's like learning your multiplication tables. Once you've done it enough, you won't even use math anymore. It'll just be memorized.

It's basically just doubling the number. I usually do up to 65536 (2^16) in my head. I also write them down as well. 2 4 8 16 32 64 128 256 1024 2048 4096 8192 16384 32768 65536. Just think of RAM :d.
Agreed. Every good computer nerd should know the power of two to at least 2^16. I followed the progress of video RAM from the 8-bit color days, so I got it all before I ever got into networking.
 
Once you have learn Subneting; then try to understand Superneting.........
It will gives you headache :D
 
Back
Top