Free Certification Practice Tests and Study Guides
Join Us! | Login | Help

Subnetting Confusion

SubnettingWe recently released an article entitled Battling Incorrect Information in which we mentioned some of the confusion regarding subnetting. After receiving some requests for clarification, we decided to put together an article on the subject since it will affect what you need to know to pass many of the various certification exams that will test you on TCP/IP.

First, lets take a look at where the confusion lies. When subnetting a network the IP protocol and IP structure have always defined all subnets, but legacy software usually does not recognize the subnets created that have all 1's or all 0's as defined by RFC 950. This is why traditional subnetting tables utilize the 2N-2 (where N is the number of bits added to the mask for subnetting)formula to document the list of usable subnets. For instance, if you want to use 1 bit for subnetting, by using this formula you end up with 0 usable subnets. This is because with one bit, it is either on (1) or off (0). Since older software did not support this, there are 0 usable subnets. With 2 bits for the subnet, the formula yields 2 usable subnets 10 and 01 (00 and 11 were not supported by the software). This essentially wasted 2 whole subnets and their associated host addresses.

Modern software makes this waste of IP addresses obsolete and is able to support subnets utilizing all 1's or 0's in the mask as defined by RFC 1878. For example, traditionally you would not be able to use the following subnet structure which is now supported:

Subnet the class C network into 2 subnets

Subnet mask: (1 bit subnet)

Subnet 1:
Hosts 1 : -
Broadcast 1:

Subnet 2:
Hosts 2: -
Broadcast 2:

This applies across all classful ranges and remakes the formula from 2N-2 to just 2N (where N is the number of bits added to the mask for subnetting).

Here is a link to the RFC referencing the use of all definable networks: Click here

Also here is a link to a modern table on Cisco's site where a class B range is subnetted and all definable subnets are used(Note the "Subnet Portion of Last Octet (Binary)" column in table A-1 which shows the subnet bits): Click here

So which way do you need to learn for the exams? It depends on the exam you are taking. Below we will outline what we know about some of the certifications.
  • MCSE - While Windows 2000 is RFC 1878 compliant, the MCSE exams will still be testing using the 2N-2 method.
  • Cisco - With the release of their 2.0 track, they are now testing on the 2N method.
  • CompTIA - CompTIA exams do not test this in depth so you will probably not run into any questions on their exams that will be affected by this.
Hopefully that clears up some of the confusion. We have updated the study guides on our site to reflect this new information.