IP Addressing — IPv4, IPv6, Classes, Address Types & CIDR

Topics: IPv4 | IPv6 | Classes | Public IP | Private IP | Loopback | APIPA | Static IP | Dynamic IP | CIDR

Table of Contents

1. What is an IP Address?
2. IPv4 — Structure & Binary Basics
3. Classful Addressing (Classes A, B, C, D, E)
4. CIDR — Classless Inter-Domain Routing
5. Public vs Private IP Addresses
6. Loopback Address (127.0.0.1)
7. APIPA — Automatic Private IP Addressing (169.254.x.x)
8. Static IP vs Dynamic IP (DHCP)
9. IPv6 — Internet Protocol Version 6
10. Quick Reference Tables

What is an IP Address?

An IP Address (Internet Protocol Address) is a unique numeric identifier assigned to every device connected to a network that uses the Internet Protocol. It serves two functions:

Two Versions

Two standards for IP addressing — IPv4 (32-bit, current) and IPv6 (128-bit, next generation).

IPv4 — Structure & Binary Basics

A 32-bit addressing system using dotted-decimal format (4 octets, each 0–255) to uniquely identify devices on a network.

IPv4 Format

Four 8-bit octets separated by dots, written in decimal (0–255 per octet).

An IPv4 address is a 32-bit number written as 4 decimal octets separated by dots:

192 . 168 . 1 . 10
 │     │    │   └── 4th octet (8 bits → 0–255)
 │     │    └────── 3rd octet (8 bits → 0–255)
 │     └─────────── 2nd octet (8 bits → 0–255)
 └───────────────── 1st octet (8 bits → 0–255)

Total: 32 bits = 4 bytes = 4,294,967,296 possible addresses.

Binary for IP

Computers process IPs in binary. Each octet is 8 bits:

Bit position:  128  64  32  16   8   4   2   1
Binary:          1   0   1   0   1   0   0   0  = 168
                 0   0   0   0   1   0   1   0  =  10

Decimal → Binary: Subtract the largest power of 2 each time.

168 → 128 (1), 40 left → 64 (0), 40 → 32 (1), 8 left → 16 (0),
      8 → 8 (1), 0 left → 4 (0), 2 (0), 1 (0)
Result: 10101000

Binary → Decimal: Add up positions where bit = 1.

10101000 → 128 + 32 + 8 = 168

Key Binary Values to Memorize

The 8 most common binary values used in subnet masking.

Classful Addressing (Classes A, B, C, D, E)

The original IP allocation system (now obsolete) that divided IPv4 addresses into fixed-size classes based on the first few bits.

How to Identify the Class

The first octet value determines the class — 1–126 = A, 128–191 = B, 192–223 = C, 224–239 = D, 240–255 = E.

Look at the first octet (first 8 bits):

Note: 127.x.x.x is reserved for loopback, not part of Class A.

Class A (1.0.0.0 – 126.255.255.255)

Network.Host.Host.Host — 126 networks with 16 million hosts each, used by ISPs and large enterprises.

Structure:  Network . Host . Host . Host
Networks:   126
Hosts/net:  2²⁴ − 2 = 16,777,214
Used by:    ISPs, tech giants, AWS/Azure/GCP VPCs
Example:    10.0.0.1 → network = 10, host = 0.0.1

Class B (128.0.0.0 – 191.255.255.255)

Network.Network.Host.Host — 16,384 networks with 65,534 hosts each, used by medium-large organizations.

Structure:  Network . Network . Host . Host
Networks:   2¹⁴ = 16,384
Hosts/net:  2¹⁶ − 2 = 65,534
Used by:    Medium-large organizations, universities
Example:    172.16.0.1 → network = 172.16, host = 0.1

Class C (192.0.0.0 – 223.255.255.255)

Network.Network.Network.Host — 2 million networks with 254 hosts each, used by small organizations and home networks.

Structure:  Network . Network . Network . Host
Networks:   2²¹ = 2,097,152
Hosts/net:  2⁸ − 2 = 254
Used by:    Small organizations, home networks
Example:    192.168.1.10 → network = 192.168.1, host = 10

Class D (224.0.0.0 – 239.255.255.255)

Not for host addressing — used for multicast (one-to-many communication).

Used for multicast (one-to-many), not host addressing:

• 224.0.0.1 — All hosts on subnet
• 224.0.0.5 — All OSPF routers
• 224.0.0.251 — mDNS (Bonjour/AirPrint)

Class E (240.0.0.0 – 255.255.255.255)

Reserved for experimental / research use only.

Comparison

A quick reference comparing the three main classes by size, capacity, and typical use.

CIDR — Classless Inter-Domain Routing

A modern IP addressing system (1993) that replaces rigid classes with flexible prefix lengths to prevent address waste.

Why CIDR?

CIDR solved the problem of wasted IP addresses by allowing any prefix length instead of being locked into /8, /16, /24.

The classful system was wasteful. A company needing 300 hosts had only two options:

Class C (/24):  254 hosts  → Too small
Class B (/16):  65,534 hosts → 65,234 wasted!

CIDR (pronounced "cider") replaced classes in 1993. It allows any prefix length, not just /8, /16, /24.

CIDR Notation Format

Written as IP/Prefix (e.g., 192.168.1.0/24) — the number after / counts the network bits.

192.168.1.0 / 24
              │
              └── Number of bits reserved for the network

The /24 is shorthand for subnet mask 255.255.255.0. The number after / simply counts how many leading 1s are in the mask.

From CIDR to Subnet Mask

Convert CIDR to mask by writing N consecutive 1s followed by (32-N) zeros, then convert each 8-bit group to decimal.

/24 = 24 ones + 8 zeros → 11111111.11111111.11111111.00000000 → 255.255.255.0
/28 = 28 ones + 4 zeros → 11111111.11111111.11111111.11110000 → 255.255.255.240
/22 = 22 ones + 10 zeros → 11111111.11111111.11111100.00000000 → 255.255.252.0

From Subnet Mask to CIDR

Count the binary 1s in the mask — each 255 = 8 bits, then add the bits from the remaining octet.

255.255.255.0   →  24 ones → /24
255.255.255.240 →  28 ones → /28
255.255.252.0   →  22 ones → /22

Conversion Table (Most Used)

The 7 most common CIDR values — memorize these for subnetting.

CIDR vs Classful: With CIDR, that company needing 300 hosts can use /23 (510 addresses) — no waste.

Public vs Private IP Addresses

Public IPs are globally unique and routable on the internet; Private IPs (RFC 1918) are reused internally and require NAT to reach the internet.

Public IPs

A globally unique IP address routable on the public internet, assigned by IANA through ISPs.

• Routable on the internet
• Globally unique — no two devices share the same public IP
• Assigned by IANA → RIRs → ISPs
• Usually leased (static public IPs may cost extra)

Examples: 8.8.8.8 (Google DNS), 1.1.1.1 (Cloudflare), 4.2.2.2

Private IPs (RFC 1918)

IP addresses reserved for internal use — not routable on the internet and can be reused across different private networks.

• Not routable on the internet
• Reusable across private networks (your 192.168.1.x ≠ your neighbor's)
• Must use NAT (Network Address Translation) to access the internet

The Three Private Ranges

Three blocks reserved by RFC 1918 — 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16 — covering Class A, B, and C sizes.

How NAT Works

NAT translates private IPs to a public IP so multiple devices can share one internet connection.

Laptop (192.168.1.10) ──→ Router ──→ Internet (203.0.113.50)
        (private)              ↑ NAT translates private → public

The router keeps a mapping table so response packets find their way back to the correct device.

Loopback Address (127.0.0.1)

A special IP address used by a device to send network traffic to itself — data never leaves the machine.

What Is It?

A loopback address sends traffic to the device itself — the data never leaves the machine. It's handled entirely by the OS network stack.

IPv4 Loopback

Range 127.0.0.0/8, standard address 127.0.0.1, also known as "localhost."

• Range: 127.0.0.0/8
• Standard address: 127.0.0.1
• Hostname: "localhost"

ping 127.0.0.1  → Tests your TCP/IP stack (software loopback)
                  Does NOT test your network cable or NIC

Why It's Useful

Loopback is used for testing TCP/IP, running local development servers, and security isolation.

IPv6 Loopback

The IPv6 equivalent of 127.0.0.1 — address ::1/128.

• Address: ::1/128

Common Localhost Ports

Standard port numbers used with localhost for common development services.

Port 80:     http://localhost
Port 443:    https://localhost
Port 3000:   React / Node.js dev server
Port 3306:   MySQL
Port 5432:   PostgreSQL

APIPA — Automatic Private IP Addressing (169.254.x.x)

A fallback feature that auto-assigns a link-local IP (169.254.x.x) when a device cannot reach a DHCP server.

What Is It?

APIPA is a fallback feature (Windows and other OSes). When a device is set to use DHCP but no DHCP server responds, it auto-assigns itself a link-local IP.

APIPA Range

169.254.0.1 – 169.254.255.254 with /16 mask, limited to the local subnet only.

Range:  169.254.0.1 – 169.254.255.254
Mask:   255.255.0.0 (/16)
Limit:  Link-local only — never routed, no internet access

When APIPA Happens

The boot sequence — DHCP Discover gets no response → device assigns itself an APIPA address.

Device boots → needs IP → DHCP Discover → no response → APIPA: 169.254.x.x
                                                        └── Internet broken

If You See 169.254.x.x

A troubleshooting guide — common causes and what to check when your device shows an APIPA address.

 Possible causes:
   • Router / DHCP server is down
   • Network cable unplugged
   • Wi-Fi disconnected / wrong password
   • Faulty network adapter

 Check first:
   1. Is the router powered on?
   2. Is the cable plugged in?
   3. Restart the network adapter

Static IP vs Dynamic IP (DHCP)

Static IPs are manually configured and never change; Dynamic IPs are automatically assigned by DHCP and may change over time.

Static IP

A fixed IP address manually configured on a device — predictable but requires manual management.

Manually configured — never changes.

Example config:
  IP:        192.168.1.100
  Mask:      255.255.255.0
  Gateway:   192.168.1.1

Dynamic IP (via DHCP)

An IP address automatically assigned by a DHCP server — no configuration needed but the IP may change.

Automatically assigned by a DHCP server — may change over time.

DHCP in Action (DORA)

A 4-step process — Discover, Offer, Request, Acknowledge — by which a device obtains an IP from a DHCP server.

Client                     DHCP Server
  │                              │
  │── DISCOVER (Anyone there?) ──→│
  │←── OFFER (Use 192.168.1.10) ──│
  │── REQUEST (I'll take it!) ──→│
  │←── ACK (Confirmed!) ─────────│
  │                              │
  IP: 192.168.1.10  Lease: 24 hours

DHCP Reservation (Best of Both)

A DHCP server setting that always assigns the same IP to a specific device (by MAC address) — combines static stability with dynamic ease.

A DHCP reservation ties a specific IP to a device's MAC address on the DHCP server:

Printer (MAC: AA:BB:CC:00:11:22) → always gets 192.168.1.50

• Device still uses DHCP (no manual config)
• But always gets the same IP (like static)
• Managed centrally on the router/DHCP server

When to Use Which

A guide to choosing static vs dynamic IP based on the device's role and requirements.

IPv6 — Internet Protocol Version 6

The next-generation IP protocol using 128-bit addresses (written in hex) to solve IPv4 address exhaustion.

Why IPv6?

IPv4 has ~4.3 billion addresses — exhausted in 2011. IPv6 provides 340 undecillion addresses (2¹²⁸).

IPv6 Format

128 bits → 8 groups of 16 bits each, written in hexadecimal and separated by colons.

128 bits → 8 groups of 16 bits, written in hexadecimal, colon-separated

Full:    2001:0db8:85a3:0000:0000:8a2e:0370:7334
         │      │      │      │      │      │      │      │
         └──1───┘ └──2───┘ └──3───┘ └──4───┘ └──5───┘ └──6───┘ └──7───┘ └──8───┘
         Each group = 16 bits = 4 hex digits

Shortening Rules

Two rules to shorten IPv6 addresses — omit leading zeros, and replace one contiguous zero-run with ::.

Rule 1 — Omit leading zeros in any group:

2001:0db8:0000:0000:0000:8a2e:0370:7334
→ 2001:db8:0:0:0:8a2e:370:7334

Rule 2 — Replace one contiguous run of zero-groups with :: (use only once):

→ 2001:db8::8a2e:370:7334

More examples:

Full:     2001:0000:0000:0000:0000:0000:0000:0001
Short:    2001::1

Full:     fe80:0000:0000:0000:0213:72ff:fe9a:a5b0
Short:    fe80::213:72ff:fe9a:a5b0

IPv6 Address Types

Five main types — Global Unicast (internet), Unique Local (private), Link-Local (subnet), Loopback (localhost), Multicast (one-to-many).

No broadcast in IPv6 — multicast replaces it.

IPv6 vs IPv4 Quick Comparison

Side-by-side comparison of key differences between IPv4 and IPv6.

Dual Stack

Running both IPv4 and IPv6 simultaneously — the OS tries IPv6 first and falls back to IPv4.

Most modern networks run IPv4 + IPv6 simultaneously. Your OS tries IPv6 first, falls back to IPv4.

Quick Reference Tables

All Special IPv4 Addresses

A complete list of reserved IPv4 ranges and their purposes.

Fast Class Identification

Quick reference to identify IP class by the first octet value.