The network part of an IP address is called the network address. The network address identifies a unique network segment. A network administrator can divide a network address into subnets so that broadcast packets are transmitted within a single subnet.
From the perspective of address allocation, subnets are supplements to network addresses. Only the net-id is assigned so that IP addresses can be used flexibly when an enterprise applies for IP addresses. The specific host-ids are assigned by the enterprise as long as there is no repetition of host IDs in the Intranet.
When hosts are widely scattered on a network, you can divide the internal host-ids into many subnets. Through the subnet classification, the entire network can be divided into smaller networks.
Subnets on an enterprise network are invisible outside the enterprise. When an external packet enters the enterprise network, the internal devices select the routes based on the subnet ID. The devices then forward the packet to the destination host.
Figure 1 shows the subnetting of a Class B IP address. The subnet mask consists of a string of continuous 1s and 0s. The 1s correspond to the net ID field and the subnet ID field. The 0s correspond to the host ID field.
Figure 1 Subnetting of a Class B address
After performing an AND operation on the 32-bit IP address and the corresponding subnet mask, you can get the net ID of an IP address. If the IP address is 10.1.1.2 and the subnet mask is 255.255.0.0, you can get 10.1.0.0 as the network address after performing the AND operation on the IP address and the corresponding subnet mask.
Subnetting reduces the available IP addresses for hosts. For example, an IP address of Class B originally can accommodate 65534 host IDs. After a 6-bit subnet field is classified, there can be a maximum of 64 subnets. Each subnet has a 10-bit host ID, which means each subnet has a maximum of 1022 (210-2, except the host IDs with all 1s and all 0s) host IDs. Therefore, there are 65408 (64 x 1022 = 65408) host IDs, 126 less than the number of IDs before subnet classification.
If an enterprise does not divide its network into subnets, the subnet mask is the default value. The number of 1s in the subnet mask indicates the net ID length. Therefore, the default values of the subnet mask for Class A, Class B and Class C IP addresses are 255.0.0.0, 255.255.0.0, and 255.255.255.0 respectively.
During subnetting and IP address planning, consider the following rules to implement reasonable and efficient network planning:
To divide network in hierarchy, consider geographic and service factors so that subnetting is subject to network hierarchies in top-down mode. In this manner, networks are effectively managed and routing tables are simplified. In most cases:
A network consisting of a backbone network and a MAN is divided into flattened subnets.
An administrative network is divided into multi-level subnets.
Consecutive addresses facilitate routing aggregation on a hierarchical network, which greatly reduces the number of routing tables and improves route searching efficiency. When allocating IP addresses, note the following issues:
Allocate consecutive IP addresses to each area.
Allocate consecutive IP addresses to devices that have the same service and function.
Do not allocate the same address to different areas even if the MPLS/VPN technique, which supports address overlapping, is used.
When allocating addresses, reserve certain addresses in each hierarchy. In this manner, consecutive addresses can be allocated to an expanded network, implementing long-term network planning.
A backbone network must have enough consecutive addresses for independent ASs and further network expansion.
When planning subnets, fully use address resources as follows to ensure that are sufficient IP addresses for hosts:
Use variable-length subnet masking (VLSM) to fully and properly use address resources.
Considering the routing mechanism of networks to fully use the IP address spaces that have been allocated for better IP address utilization.
Devices that have similar functions should be allocated IP addresses of the same type. IP address allocation complies with the following rules:
High-end switchs, IP telephony gateways, IP telephony gatekeepers, Internet servers, firewalls, and edge or access switchs should be allocated public network IP addresses. Devices running VPN services can be allocated private addresses within VPNs.
Loopback interfaces, which are device management interfaces, should be allocated a section of consecutive IP addresses with 32-bit masks.
Interfaces connecting devices should be allocated a section of consecutive IP addresses with 30-bit masks.
