An Overview of TCP/IP Internetworking

The Internet evolved from the ARPANET , an early wide-area network developed by the US military for military and academic communication. Amongst the design aims was the ability to provide a scalable internetwork of heterogeneous networks. The result was the TCP/IP-based Internet, which is based on several fundamental principles:

• Hosts in the Internet are identified by unique (32-bit) addresses  assigned by a number of central authorities. The address consists of a network part and a host part. All hosts on the same network have the same network part to their address, but unique host parts. Clearly, a host which is connected to more than one network (called a multi-homed host) has more than one address. A drawback of this approach is that moving hosts between networks requires their addresses to be changed. On the other hand, the approach has numerous advantages that more than compensate for this.

• Interconnection among networks is provided by multi-homed hosts called gateways  or routers  (a firewall is a special type of gateway). A gateway relays network traffic between two or more networks. These networks may have further gateways, resulting in further relaying. To reduce the amount of information needed by gateways, decisions about where to relay traffic (called routing decisions)  are based on the network part of the destination address; the host part is only used to route traffic once it reaches the destination network.

• All networks are treated equally. LANs, WANs and point-to-point links are each separate networks and are treated in a uniform fashion.

The network protocol used for transferring traffic across the Internet is IP (Internet Protocol), and the 32-bit addresses are known as IP addresses.