Subnetting
How IP Addressing Became Less Wasteful

Chapter 2

To help break up this unwieldy mess, CIDR was introduced.

CIDR was developed from the same principles used in normal subnetting. Instead of just seeing an octet as full or empty, CIDR forces routers to look at the octets of the subnet mask itself to determine how many addresses are available. The notation used for CIDR based subnets is the first IP address in the block followed by a forward slash and the number of 'full' bits in the subnet mask. For the Class A listed above, the CIDR formatted notation would be 198.0.0.0/8. A Class B is /16 and a /24 is a Class C.

CIDR blocks still have subnet masks associated with them. A block of 32 addresses, for instance, would be a /26. Its subnet mask would be 255.255.255.224. Confused? You don't have to be. There are some shortcuts that make it a lot easier to calculate things.

The easiest way to determine what the subnet mask looks like is to start with the classful boundary. So determine first if the block is smaller or larger than a Class C. If the block is smaller, subtract the number of IP addresses in the block from 256 and that will be the last octet of the subnet mask. Likewise, if the block is larger than a Class C and smaller than a Class B, take the number of Class C's that make up the block and subtract that number from 256. For instance, a /20 is the equivalent of 8 Class C's. 256 minus 8 is 248. So the subnet mask for a /20 would be 255.255.248.0.

Another aspect of IP addresses which confuse a lot of people are the way that they are separated into the four 'octets' and why that is done. The first thing that needs to be understood is that IP addresses are binary numbers written into decimal form for easier comprehension by people.

Each octet is 8 bits of data. Each bit is either a 1 or a 0. The first bit is 2º which is equal to 1. The second bit is 2¹ which equals 2. Third is 2², or 4. This continues until the eighth bit, which is 128. Adding the value of all of these bits together, you get 255. But there are 256 possible answers, since 0 is the number that the addresses begin on. So we are left with 256 numeric possibilities between 0 and 255 for each octet.

IPv6 uses the same sort of format, but instead of 4 octets, there are 16. Why the huge increase? Though IPv6 addresses are binary, just like IPv4 addresses, they are written out in hexadecimal form. Hexadecimal (or 'hex') is base 16, unlike decimal, which is base 10, or binary, which is base 2. So instead of 4 binary octets, IPv6 can be written out as 4 hex octets. This is mostly for convenience, since people are already used to the four octet format.

The reason for IPv4's octet system is the classful boundaries that were discussed above. As that reasoning is no longer important (thanks to subnetting and CIDR), the team developing IPv6 didn't have to worry about it. It's all a stylistic thing, now.

Another question that a lot of people have is 'What happened to IPv5?' In truth, I have never been able to find a good answer to this. As far as I can tell, there had to have been a version 5 (as most of the engineers developing such things are too detail oriented to simply skip a number like that), but it apparently never caught on. Perhaps the initial concept didn't allow for the growth patterns predicted by the time it was ready and they junked it.

Once IPv6 is implemented, there will likely never be an IP shortage again. Instead of the 4.3 billion addresses available in IPv4, IPv6 will have over 300 undecillion addresses. And what is an undecillion other than a very silly sounding word?? Imagine a 1 with 36 zeroes behind it. That means more IP addresses than if you multiplied Bill Gates' net worth by Donald Trump's.

Back to Chapter - 1