Subnetting & IP Addressing

By Alec Gauldin, Principal Network Engineer · Updated July 2026

New to this? Start with the explainer: What Is Subnetting? Subnetting Explained

Subnetting and IP addressing sit in the Network Fundamentals area of the CCNA, and they are the one skill that surfaces everywhere else: you cannot summarize a route, write an access list, bring up OSPF, or hand-address a topology without doing this math cleanly and fast. Exact exam weightings shift between blueprint revisions, so treat any percentage as secondary. The practical reality is simpler: a candidate who is slow or shaky here loses time on every other question, so the goal of this hub is fluency, not memorization.

Start with the subnet mask itself, which draws the line between the network bits and the host bits of a 32-bit address. From the host bits you get the size of each subnet, where usable hosts = 2^(host bits) - 2, because you subtract the network address and the broadcast address. The two exceptions are a /31, which carries two usable addresses for point-to-point links (RFC 3021), and a /32, a single-host route. The block-size (or 'magic number') method - 256 minus the interesting mask octet - gives you the subnet increment, so you can find network and broadcast boundaries by hand without a calculator.

Once fixed-length subnetting is comfortable, VLSM is the next layer: applying different mask lengths inside the same major network so a 50-host LAN takes a /26 while a two-host WAN link takes a /30 or /31, instead of wasting a /24 on each. The 'Subnet with VLSM (No Calculator)' guide walks that method end to end. Closely related is the wildcard mask - the bit-inverse of the subnet mask, or 255.255.255.255 minus the mask - which is what OSPF network statements and ACLs actually match on. A /24 becomes wildcard 0.0.0.255, and getting the inverse wrong is one of the most common reasons an OSPF adjacency or an ACL silently fails.

IPv6 addressing reframes the same ideas at 128 bits, and the 'Configure IPv6 Addressing on Cisco' guide covers the address types you must recognize: global unicast (2000::/3), link-local (fe80::/10, present on every IPv6-enabled interface), unique local, and multicast (ff00::/8). There is no broadcast in IPv6 - its role is taken over by multicast. Loopbacks tie both worlds together: a loopback is an always-up virtual interface used as a stable router ID and test target, and the 'Configure a Loopback Interface' guide shows how OSPF advertises a loopback as a /32 host route by default unless you change its network type.

The way to master this topic is a loop: understand the why first, keep the Subnetting Cheat Sheet at hand for the mask-to-block-size table and powers of two, then build and grade the labs until the arithmetic is muscle memory. The 'Best Subnet Calculator' guide is deliberate about this - a calculator is fine for checking your work, but the exam is timed and gives you none, so you still subnet by hand. Reach for the cheat sheet while you are learning, and lean on it less as the labs make the block-size method second nature.

Work the pieces in the order they build on each other: masks and host math, then the block-size shortcut, then VLSM, then wildcard masks, and finally IPv6 and loopbacks. Each graded lab in the bundle forces a real config on live Cisco IOS and checks it, so a wrong mask or a bad wildcard fails the grade the same way it would fail in production - which is exactly the feedback that turns memorized rules into a skill you can perform under time pressure.