1.4IPv4 addressing and subnet design

Subnet design is not a plug-into-a-formula calculation—it is genuine practical judgment: distributing a limited address space without waste, sized to each department's or floor's device count. CCNA strongly expects you to hand-calculate subnets quickly on paper. This section builds an understanding of how CIDR/VLSM work, then practices deriving a subnet layout from real requirements (per-department device counts) and verifying it by hand.

1.4.1IPv4 structure and private/public

• An IPv4 address is 32 bits, written in decimal as four 8-bit octets (e.g., 192.168.1.10). It comprises a network portion plus a host portion, with the boundary shown by the subnet mask (255.255.255.0 = /24). The historical Classes A/B/C are now superseded by CIDR (classless), which cuts the boundary freely via prefix length.
• Private addresses are internal-use ranges not routed directly on the internet: 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16. When communicating externally, NAT translates them to a public address. Public addresses are globally unique on the internet and assigned by registries.

1.4.2Subnet masks and host counts

• The number of addresses assignable to hosts in a subnet is 2^(host bits) - 2. Subtract 2 because the all-0 network address and the all-1 broadcast address cannot be assigned. Example: a /24 has 8 host bits, so 2^8 - 2 = 254 hosts.
• Extending the prefix by one bit (e.g., /24 -> /25) halves the subnet and its host count each time. The basic procedure is: from the required count, find the smallest n (host bits) satisfying 2^n - 2 >= required hosts, then derive the prefix length 32 - n.

1.4.3Variable-length splitting with VLSM

• VLSM (Variable Length Subnet Mask) splits a single address block into subnets of different lengths sized to each one's required host count. Unlike fixed-length splitting that gives every subnet the same size, it minimizes wasted addresses. The procedural tip is to carve out subnets starting from the largest requirement first.

Suppose that, as a network designer, you are given the requirement to distribute the 192.168.10.0/24 (256 addresses) allocated to a site across Sales (100 hosts needed), Development (50), and General Affairs (20) using VLSM without waste. The tip is largest-first. Sales' 100 hosts need the smallest n satisfying 2^n - 2 >= 100, which is n=7 (2^7 - 2 = 126 >= 100; n=6 gives only 62), so a 7-bit host portion = /25 (255.255.255.128). Reserve 192.168.10.0/25 (.0 to .127, assignable .1 to .126, 126 addresses) for Sales; 126 >= 100 satisfies it. From the remaining 192.168.10.128/25, carve Development's 50: the smallest n for 2^n - 2 >= 50 is n=6 (2^6 - 2 = 62 >= 50), so /26 (255.255.255.192): 192.168.10.128/26 (.128 to .191, 62 assignable). Finally, carve General Affairs' 20 from the remaining 192.168.10.192/26: the smallest n for 2^n - 2 >= 20 is n=5 (2^5 - 2 = 30 >= 20), so /27 (255.255.255.224): 192.168.10.192/27 (.192 to .223, 30 assignable). This satisfies the requirement while preserving the remaining 192.168.10.224/27 (30 addresses) for future expansion. A fixed-length /26 (62 hosts) for every department would break because Sales' 100 hosts do not fit; conversely, giving every department Sales-sized /25 would hand General Affairs 126 addresses, a huge waste. Cutting the smallest sufficient size per requirement, largest-first—that is the essence of VLSM, an address-resource allocation judgment rather than mere arithmetic.

1.4.4Section summary

• Private ranges are 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16; external communication is translated to public via NAT
• Assignable host count is 2^(host bits) - 2 (excluding all-0/all-1); derive the prefix from the smallest n with 2^n - 2 >= need
• VLSM carves variable lengths largest-first to minimize waste; always verify subnet math by hand