In network planning, the design of the IP address scheme is critical. The good IP address scheme can not only reduce network load, but also lay a good foundation for future network extensions. This article discusses some discussions on the IP address scheme design in the network plan, and gives an IP address scheme design instance. First, the IP address and subnet mask IP address is used to identify unique machines on the network. According to the definition of the RFC791, the IP address consists of 32-bit binary number (four bytes), indicating that the 12-bit decimal number (xxx.xxx.xxx.xxx) each 3-digit representation is divided into a circular point. 8-bit binary (one byte). Since the maximum number of 1 bytes can be represented by 255, each byte can contain a value between 0 to 255 in the IP address. However, 0 and 255 have special meanings, 255 represents broadcast addresses: IP address 0 0 is used to specify network address numbers (if 0 at the address end) or node address (starting if 0). For example, 192? 168? 32? 0 refers to a computer 192.168.32.0, and 0.0.0.62 refers to a computer on the network's address 62. Depending on the IP address, the IP address is divided into three types, Class C, Class C, Class C, Class A, Class C, depending on the IP address. Class A is used for oversized networks (million nodes), Class B is used for medium-scale networks (thousands of nodes), Class C is used for small networks (up to 254 nodes). The A class address represents the network address with the first byte, and the last three words represent the node address. Class B address represents the network address by the first two bytes, and the latter two bytes represent the node address. Class C addresses use the first three bytes to represent the network address, and the fourth byte represents the node address. The network device determines the network type according to the first byte of the IP address. The first binary position of the first byte of the Class A network is 0; the first two binary locations of the first byte of the Class B network are 10; the first three binary position of the first byte of Class C network is 110 See Table 2. Changes to decimal visible Class A network address from 1 to 127, Class B network address from 128 ~ 191, Class C network address from 192 ~ 223.224 ~ 239 is sometimes referred to as D class, and 239 network numbers are retained. The subnet mask is used to find the network and node address portion in the IP address. The subnet mask length is 32 bits, and 1 represents the network portion, 0 represents the node address portion. The subnet mask of Class A, Class B, and Class C is shown in Table 3, such as a node IP address is 192.168.202.195, subnet mask 255.255.255.0, indicating that its network address is 192.168.202, and the node address is 195. Sometimes in order to facilitate network management, you need to divide the network into a number of network segments. To this end, it is necessary to break the traditional 8-bit boundary, "grab" from the node address space as the network address. Specifically, establish a subnet mask requires the following two steps: 1. Determine the number of network segments running IP 2, determine the subnet mask First, determine the number of network segments running IP. For example, there are five network segments on your network, but only the users on the three network segments accesses the Internet, and only the three network segments need to configure IP. After determining the number of IP network segments, then determine that several digits from the node address space can create a sub-network number for each network segment. The method is to calculate the combined value of these bits. For example, take two bits, there are four combinations (00, 01, 10, 11), and take three combinations (000, 001, 010, 011, 100, 101, 110, 111). It should be noted that the combination of all 0 and all 1 must be removed in these groups.