This question already has an answer here:

I'm currently learning about IPv4 and subnetting. As far as I understand, if we want to create 4 small isolated networks with the size 50, we can use subnetting.

If we subnet a Class-C IPv4 network( into 4 small networks with the same size (0~63, 64~127, 128~191, 192~255), these small networks are still essential belonged to the big logical network ( even though we assign subnet masks to them.

If I understand correctly, they can communicate with each other since they share the same network portion. So, in practice, how can we isolate these small networks?


marked as duplicate by Craig Watson, Khaled, kasperd, Rob Moir, Jenny D Oct 21 '15 at 12:52

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 3
    Note that network classes aren't used anymore, so there really isn't such a thing as a Class-C network. You probably mean a network with a netmask of /24, which you're splitting up into networks with a /26 netmask. – wurtel Oct 21 '15 at 8:01
  • I might be studying some old stuff, but why and when the Class C network be deprecated? – hackjutsu Oct 21 '15 at 20:23
  • It was deprecated together with class A and class B, as experience showed that those network sizes were impractical; you don't want a broadcast domain containing a potential 16777214 hosts.Dividing networks up with arbitrary netmasks such as /22, /28 etc. makes much more efficient use of network space, and gives more manageable networks sizes. – wurtel Oct 22 '15 at 12:12
  • I see... So what about the concept of "network portion", which used to mean the first decimal of Class A, first two of Class B and first three of Class C? Is this concept also replaced by the netmask number such as /22, /28? – hackjutsu Oct 22 '15 at 17:28
  • 1
    Yes. What was class A is now shown as /8, B is /16, C is /24. You need to keep in mind that the four number parts in an IP(v4) address are simply representations of 4 sets of 8 bits (ie. the 4 bytes) of a 32 bit number. You can use 3232235777 instead of ... $ ping 3232235777 shows PING 3232235777 ( 56(84) bytes of data. – wurtel Oct 23 '15 at 7:54

If I understand correctly, they can communicate with each other since they share the same network portion.

This part is wrong. The subnets will be their own segment entirely, and will not be able to communicate with eachother - they will only be able to "see" other devices on their own localised smaller subnet.

For example, will have a network address of and broadcast address of Any devices will need to sit between and

In addition to this they will need to be separated via either logically separated VLANs or physically separated network hardware. If you need the networks to communicate, you will need to use layer 3 routing.

  • "they will only be able to "see" other devices on their own localized smaller subnet." Does this mean that the clients can see each other if they sit in same network portion even though they are in different subnets? – hackjutsu Oct 22 '15 at 21:01
  • I think you're confusing terminology. Devices can only communicate within their own subnet. The "network portion" you refer to is completely superfluous and makes no difference whatsoever. – Craig Watson Oct 22 '15 at 21:34

Your question its not quite right "How to isolate networks when subnetting?" so I will try to clarify things for you .

You have an IP and a Subnet Mask. The IP is the exact interface address, and the subnet mask extract from that IP the network address, for example.

     type      |           Network            HOST    | Hexadecimal
IP addreses    | 11000000.10101000.00000101.00000010  |
Subnet mask    | 11111111.11111111.11111111.00000000  |
Network address| 11000000.10101000.00000101.00000000  |

In order to separate("isolate") your networks, you need different network addresses :

(the 24 stands for 24 bits mask - the binary ones(1))

In order to make those networks communicating to each other, you need a Router that has an interface(physical or virtual) for each network.

Dont get confused by the numbers, look at the binary table! Numbers are separate for better understanding from us(human beings). Look at the binary and the HOST bits are the amount of hosts per network.

So about the question - You separate networks with different Subnet masks

Reference for more detailed explenation - https://en.wikipedia.org/wiki/Subnetwork

Not the answer you're looking for? Browse other questions tagged or ask your own question.