Illustrated TCP/IP by Matthew G. Naugle Wiley Computer Publishing, John Wiley & Sons, Inc. ISBN: 0471196568   Pub Date: 11/01/98

Chapter 69
IP Address Assignment (The Old Method)

Originally, using RFC 791 without subnetting, an organization with a complex (more than one) network topology had three choices for assigning Internet addresses:

1.  Acquire a distinct Internet network number for each cable; subnets are not used at all.

2.  Use a single network number for the entire organization, but assign host numbers in coordination with their communication requirements (flat networks segmented using bridges).

3.  Use a single network number and partition the host address space by assigning subnet numbers to the LANs (explicit subnets). Create your own but don’t advertise them to the ARPAnet. This is the most popular method.

Employing the first choice caused routing tables to grow. RFC 950 allowed for subnet addressing to take place within an autonomous system, which allowed for a site to continue to subnet its AS, but the subnets were never propagated to the Internet routing tables. Subnetting and VLSM (variable-length subnet masks, explained later) allowed for the global routing tables to stop growing exponentially and allowed sites to control their own networks as well. However, network numbers were plentiful and subnets slowed the expansion of the Internet routing tables. This was before the commercialization of the Internet in 1994.

The adverse effects of bridges in complex networks are well known. Since the bridge revolution, routers have become the mainstay of the corporate backbone. This worked well for shared environments, but technology was changing: Network attachments were becoming faster and more powerful. The bridging revolution came back as switches in that each desktop could now have its own 10-Mbps pipe. The switches build a small flat network and should be used to front end routers, thus allowing for microsegmenting but not microsubnetting.

Subnetting one network number caused the Internet routing tables to slow their growth. This worked well with Class B addresses. Class C networks forced the Internet routing tables to grow, and Class A addresses were not handed out. Also, since more than 50 percent of the businesses were small- and medium-sized businesses. Class C addresses were needed. Again, we were in a predicament. We needed a solution.