As most people know it, the Internet is a myth.
Most people picture the Internet as a single network the spans the globe and reaches into the businesses and households of everyone who connects. In reality, it was the interconnection of millions of smaller networks that created the global network now known as the Internet.
To achieve this incredible feat, Al Gore, the Advanced Research Projects Agency (ARPA), and various universities joined together to outline the rules that would permit the widest range of technologies to unite and form the information superhighway.
These rules defined the construction of the Internet.
To implement a global network while permitting decentralized administration of that network, they created several standards that are now the mainstays of modern computer networks.
These achievements included the definition of Domain Name Services (DNS), the TCP/IP protocol, and TCP/IP networks and subnetworks. This article will specifically address the basics of subnetworks.
Subnetting, or the process of creating subnetworks, is the method used to partition or divide a single network into smaller networks. We can continue to create subnets on each of these smaller networks in turn until we reach the minimum network size of two network "nodes". For each network, we must have a network identity, and for each network node, we need to have a node address. The combination of the network identity with the node address must result in a totally unique address.
With TCP/IP v4, this unique address is actually a 32-bit binary number that is normally written as a four-digit decimal number similar to 208.112.8.62. The lower boundary of this range is 0.0.0.0, whereas the upper boundary is 255.255.255.255. This permits a total of over 4 billion addresses available on the global network.
To determine the actual network identity and node address of a particular device connected to the internet, we use another number often referred to as the "network mask" or "net mask". To differentiate between the two components, the entire address is written in binary and the net mask is used to "mask" the network identity, and all other digits left over become the node address. This is easier to understand if you see it in action.
Assume that we have an address of 208.112.8.255 (not a valid address, this is used for explanation only), with a net mask of 255.255.255.0. This would be written as:
• 11010000.01110000.00001000.11111111 Unique Address
• 11111111.11111111.11111111.00000000 Net Mask
• 11010000.01110000.00001000.00000000 Network Identity
• 00000000.00000000.00000000.11111111 Node Address
Notice that in every "position" where a "1" is in the net mask, the number in the unique address is brought down into the network identity. Wherever a "0" appears in the net mask, the number is brought down into the node address. This process allows a single number to be used to decompose the full TCP/IP address into the required parts of the network ID and the node ID, and allows for networking messages to be transferred to the right computing device on the correct network.
The net mask also let the network administrators adjust the size of individual networks to have fewer people on a specific network or more people. The network used in the example above with a netmask of 255.255.255.0 would have 254 node addresses available. If we were to change the net mask to 255.255.255.240, the number of nodes available would drop to only 14. By doing this, we have effectively "subnetted" the first network into 16 separate little networks of 14 nodes each. How this works will be the subject of another article, but in a quick and non-technical summary, you take two the power of the number of empty bits in the net mask (the "0"s), and subtract 2. You subtract the 2 because this represents one for the upper boundary of the network, and the second represents the lower boundary of the network.
As the first netmask that was 255.255.255.0 (1111111.11111111.11111111.00000000) has eight empty bits, we would take 2 to the power of 8 which is 256. We subtract 2 from this total to give us a total of 254 node addresses available. In the second netmask of 255.255.255.240 (11111111.11111111.11111111.11110000), we find four empty bits, so we take 2 to the power of 4 which is 16, and subtract 2 which gives us a total of 14 node addresses available. In subnetting our network like this, we have gone from having a single network of 254 addresses to have a total of 16 networks that have 14 addresses each.
This means that an Internet Service Provider (ISP) can utilize a single range of numbers to support a number of small clients.
Though the Internet as the global network reaching its tentacles into every computer is a myth, the global backbone ties millions of small networks into a single interwoven tapestry of computing power.