IP Addressing

Updated

September 4, 2026

IP Addressing

Forwarding starts with addressing: structure, masks, subnets, summarization, and enough IPv6 to run dual-stack labs. This chapter is calculation plus lab verification—not exam trick puzzles for their own sake.

Learning goals

By the end of this chapter you can:

  • Convert between prefix length and masks fluently
  • Design subnets for a multi-tier lab without overlaps
  • Summarize contiguous blocks and predict blackholes from bad summaries
  • Configure IPv4 and basic IPv6 on Linux
  • Use documentation ranges and lab hygiene ranges deliberately

IPv4 structure

An IPv4 address is 32 bits, usually dotted decimal. A prefix address/len identifies network bits vs host bits.

Prefix Mask Hosts (approx usable)
/32 255.255.255.255 1 (host route)
/31 255.255.255.254 2 (P2P often)
/30 255.255.255.252 2 usable classic
/24 255.255.255.0 254 usable classic
/16 255.255.0.0 large
/8 255.0.0.0 huge 
# Linux
ip addr add 192.0.2.10/24 dev eth1
ip -br a
ip route

Subnetting for labs

Example campus-ish allocation from 10.20.0.0/16:

Use Block
Users site A 10.20.1.0/24
Users site B 10.20.2.0/24
Servers 10.20.10.0/24
P2P core 10.20.255.0/24 as /30s
Loopbacks 10.20.254.0/24 as /32s

Rules:

  1. No overlapping prefixes on different segments unless deliberate anycast
  2. Leave growth room
  3. Write the plan before deploy
  4. Align summaries with actual aggregation topology

Summarization

If you own 10.20.1.0/24 and 10.20.2.0/24, a summary 10.20.0.0/22 may or may not be correct depending on other contents of that /22.

Action Risk
Summary advertises space you do not have Blackhole for missing parts
Too-specific everywhere Table bloat (later IGP/BGP scale)
Forgetting more-specific local routes Traffic steers wrong

Paper drill

You have:

  • 10.0.0.0/25
  • 10.0.0.128/25

Correct tight summary? 10.0.0.0/24.

You have only 10.0.0.0/25 but advertise 10.0.0.0/24 upstream without a sink. What happens to 10.0.0.200? Blackhole toward you.

Special-use ranges (know them)

Range Role
10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16 Private (RFC 1918)
127.0.0.0/8 Loopback
169.254.0.0/16 Link-local
192.0.2.0/24, 198.51.100.0/24, 203.0.113.0/24 Documentation (RFC 5737)
224.0.0.0/4 Multicast

IPv6 essentials for dual-stack labs

IPv6 address is 128 bits; common lab prefix /64 on LANs.

ip -6 addr add 2001:db8:1::10/64 dev eth1
ip link set eth1 up
ip -6 route
ip -6 neigh
ping -6 -c 2 2001:db8:1::1
Prefix idea Example
Documentation 2001:db8::/32
Unique local fd00::/8 (with local generation rules)
Link-local fe80::/10 auto on interfaces

ND replaces ARP. Router Advertisements may autoconfigure hosts—labs often use statics for control.

Dual-stack habit

Give nodes both families when learning:

ip addr add 10.1.0.10/24 dev eth1
ip -6 addr add 2001:db8:1::10/64 dev eth1

Test both paths; do not assume v4 success means v6 works.

Linux configuration patterns

Persistent lab style

In Containerlab, prefer startup scripts or image config over manual ip addr that vanishes.

#!/bin/sh
ip addr flush dev eth1
ip addr add 10.1.0.10/24 dev eth1
ip -6 addr add 2001:db8:1::10/64 dev eth1
ip link set eth1 up
ip route add default via 10.1.0.1
ip -6 route add default via 2001:db8:1::1

Multiple addresses

ip addr add 10.1.0.10/24 dev eth1
ip addr add 10.1.0.11/24 dev eth1

Secondary addresses appear in ARP; know who answers.

Lab: addressing plan + dual-stack hosts

name: l3-addr

topology:
  nodes:
    r1:
      kind: linux
      image: alpine:3.20
      exec:
        - apk add --no-cache iproute2 iputils
        - sysctl -w net.ipv4.ip_forward=1
        - sysctl -w net.ipv6.conf.all.forwarding=1
        - ip addr add 10.1.0.1/24 dev eth1
        - ip addr add 10.2.0.1/24 dev eth2
        - ip -6 addr add 2001:db8:1::1/64 dev eth1
        - ip -6 addr add 2001:db8:2::1/64 dev eth2
        - ip link set eth1 up
        - ip link set eth2 up
    h1:
      kind: linux
      image: alpine:3.20
      exec:
        - apk add --no-cache iproute2 iputils
        - ip addr add 10.1.0.10/24 dev eth1
        - ip -6 addr add 2001:db8:1::10/64 dev eth1
        - ip link set eth1 up
        - ip route add default via 10.1.0.1
        - ip -6 route add default via 2001:db8:1::1
    h2:
      kind: linux
      image: alpine:3.20
      exec:
        - apk add --no-cache iproute2 iputils
        - ip addr add 10.2.0.10/24 dev eth1
        - ip -6 addr add 2001:db8:2::10/64 dev eth1
        - ip link set eth1 up
        - ip route add default via 10.2.0.1
        - ip -6 route add default via 2001:db8:2::1

  links:
    - endpoints: ["h1:eth1", "r1:eth1"]
    - endpoints: ["h2:eth1", "r1:eth2"]

Predict

  • h1 pings h2 v4 and v6 through r1
  • Wrong mask on h1 (e.g. /32 only) breaks on-link perception of gateway

Observe

docker exec clab-l3-addr-h1 ping -c 2 10.2.0.10
docker exec clab-l3-addr-h1 ping -6 -c 2 2001:db8:2::10
docker exec clab-l3-addr-h1 ip route get 10.2.0.10
docker exec clab-l3-addr-r1 ip -br a

Failure: overlapping subnets

Put h2 also in 10.1.0.0/24 while connected to r1 eth2 differently—create intentional overlap and journal chaos (ARP confusion, asymmetric paths). Restore clean plan.

Failure: bad summary (static preview)

On a third router path (optional), advertise/install 10.0.0.0/8 via a dead next hop while more specifics missing—observe blackhole for unused space.

Calculation practice (quick)

  1. How many /24s in a /16? → 256
  2. First/last usable in 10.0.0.0/30 classic? → 10.0.0.1 and 10.0.0.2
  3. Is 10.0.0.64/26 inside 10.0.0.0/24? → yes
  4. Can you summarize 10.1.0.0/24 and 10.2.0.0/24 as a single /23? → no (not contiguous bit boundary as one /23 covering both cleanly without others)

Work these on paper until automatic.

Host vs router addressing habits

Role Typical
Host One address + default route
Router Addresses on each L3 interface + routing table
Loopback Stable /32 or /128 for protocols later 
ip addr add 1.1.1.1/32 dev lo

Verification checklist

ip -br a
ip route
ip -6 route
ip route get <dst>
ping -c 2 <dst>
ping -6 -c 2 <dst6>

Summary

  • Prefixes define on-link vs routed behavior
  • Plan non-overlapping lab allocations; document them
  • Summaries must match reality or they blackhole
  • Dual-stack labs need explicit v6 config and tests
  • ip route get is your friend

Next: routing fundamentals—RIB/FIB, longest match, next hops, and host vs router behavior including ARP/ND.