The IP addresses are divided into three different types, based on their operational characteristics:
1. unicast IP addresses – an address of a single interface. The IP addresses of this type are used for one-to-one communication. Unicast IP addresses are used to direct packets to a specific host. Here is an example:
In the picture above you can see that the host wants to communicate with the server. It uses the (unicast) IP address of the server (192.168.0.150) to do so.
2. multicast IP addresses – used for one-to-many communication. Multicast messages are sent to IP multicast group addresses. Routers forward copies of the packet out to every interface that has hosts subscribed to that group address. Only the hosts that need to receive the message will process the packets. All other hosts on the LAN will discard them. Here is an example:
R1 has sent a multicast packet destined for 22.214.171.124. This is an RIPv2 packet, and only routers on the network should read it. R2 will receive the packet and read it. All other hosts on the LAN will discard the packet.
3. broadcast IP addresses – used to send data to all possible destinations in the broadcast domain (the one-to-everybody communication). The broadcast address for a network has all host bits on. For example, for the network 192.168.30.0 255.255.255.0 the broadcast address would be 192.168.30.255*. Also, the IP address of all 1’s (255.255.255.255) can be used for local broadcast. Here’s an example:
R1 wants to communicate with all hosts on the network and has sent a broadcast packet to the broadcast IP address of 192.168.30.255. All hosts in the same broadcast domain will receive and process the packet.
*This is because the subnet mask of 255.255.255.0 means that the last octet in the IP address represents the host bits. And 8 one’s written in decimal is 255.