Scenario

  • small business
    • 4 departments in 2 offices
  • instead of purchasing 4 separate address ranges from the internet authorities, we buy one single range (less expensive): Class C 200.15.10.0/24
    • this range can be divided up further and be assigend to the different departments
  • when subnetting, we take some of the host addresses away and give them to the network portion of the address
    • the more subnets, the fewer hosts on the individual subnets and vice versa

Calculate the number of networks

  • formula: 2^subnet-bits
    • example 1:
      • class C network uses a /28 subnet mask
      • thus, we have borrowed 4 bits from the default of /24 (28 - 24)
      • 2^4 = 16 available subnets
    • example 2:
      • class B network uses a /28 subnet mask
      • thus, we have borrowed 12 bits from the default of /16 (28 - 16)
      • 2^12 = 4,096 available subnets
  • hosts on different subnets need to go via a router if they want to communicate with each other

Calculate the number of hosts

  • formula: 2^host-bits - 2
    • subtract 2, because the network and broadcast addresses cannot be assigned to hosts
    • example:
      • class C network uses a /28 subnet mask
      • thus 4 bits left for hosts (32 - 28)
      • 2^4 - 2 = 14

Note on ip subnet-zero

  • in the past, the formula for calculating the number of networks was 2^subnet-bits - 2 (because it was not allowed to use network bits of all 0’s or 1’s)
  • however, there was no real need for that 
     # override this limitation on a router (enabled by default)
 ip subnet-zero