Consider the scale of addressing on today's Internet: billions of devices, each requiring a unique identifier, spread across every continent, connected by networks owned by thousands of different organizations, using dozens of different physical transmission technologies.

How do you create an addressing scheme that:

• Uniquely identifies billions of devices?
• Enables efficient routing through millions of routers?
• Works regardless of whether the device is connected via Ethernet, Wi-Fi, cellular, or satellite?
• Allows networks to be added, split, or reorganized without global reconfiguration?
• Permits devices to move between networks and obtain new addresses automatically?

The answer is logical addressing—a software-defined addressing scheme that operates independently of physical network hardware. Where MAC addresses are burned into network interface cards and identify the hardware, IP addresses are assigned by software and identify a device's location in the network topology.

This distinction—hardware-independent, topology-aware addressing—is what makes the global Internet possible.

To appreciate logical addressing, we must first understand what it replaces and why physical addressing alone is insufficient for global networking.

Physical Addresses (MAC Addresses):

The Data Link Layer uses Media Access Control (MAC) addresses to identify devices. These addresses have specific characteristics:

Why MAC Addresses Fail for Global Routing:

Imagine trying to build a global postal system where every address is a random 48-bit number with no relationship to geographic location. To deliver mail, every post office would need a list of every address in the world and which direction to send it. This is clearly unworkable.

MAC addresses have the same problem:

1. No Location Information: MAC 00:1A:2B:3C:4D:5E tells you nothing about where that device is. It could be in Tokyo, Texas, or the International Space Station.

2. No Aggregation: You cannot summarize MAC addresses. A router cannot say "all addresses starting with 00:1A go left" because devices with similar MAC prefixes are scattered globally.

3. Flat Routing Tables: Without aggregation, routing tables would need an entry for every device on the Internet—billions of entries, updated in real-time. This is computationally impossible.

4. Hardware Dependence: If you replace your network interface card, your identifier changes. Persistent system identification becomes impossible.

Logical addresses (IP addresses) solve the problems of physical addressing through several key properties:

1. Hierarchical Structure

IP addresses embed location information through a hierarchical structure:

An IP address consists of two conceptual parts:

• Network Portion: Identifies the network or subnet the host belongs to
• Host Portion: Identifies the specific host within that network

The division point is defined by a subnet mask (or prefix length). This hierarchy enables:

• Route Aggregation: A single routing table entry can cover thousands of hosts
• Hierarchical Delegation: IANA → Regional Registries → ISPs → Organizations → end users
• Locality: Addresses in the same network share a prefix, simplifying routing

2. Software Assignability

Unlike MAC addresses, IP addresses are assigned by software and can change:

• Static Assignment: Administrator manually configures IP address
• Dynamic Assignment: DHCP server automatically assigns addresses
• Self-Assignment: Link-local addresses (169.254.x.x, fe80::) when no DHCP available

This flexibility enables:

• Devices to move between networks and receive contextually appropriate addresses
• Network renumbering without replacing hardware
• Efficient address utilization through dynamic allocation

3. Global Uniqueness (for public addresses)

Public IP addresses are globally unique—no two devices on the public Internet have the same address (at the same time). This is enforced through a hierarchical allocation system managed by IANA and Regional Internet Registries (RIRs).

4. Scope Awareness

IP addresses have defined scopes:

• Global: Routable anywhere on the Internet (public addresses)
• Private: Valid only within an organization's network (10.x.x.x, 172.16-31.x.x, 192.168.x.x)
• Link-Local: Valid only on a single link/subnet (169.254.x.x for IPv4, fe80::/10 for IPv6)
• Loopback: Refers to the local host itself (127.0.0.1, ::1)

IPv4, the fourth version of the Internet Protocol, has been the dominant network layer protocol since the 1980s. Its addressing scheme, while showing its age, illustrates the core concepts of logical addressing.

IPv4 Address Structure:

Network and Host Division:

Every IPv4 address is conceptually divided into a network portion and a host portion. The subnet mask defines where this division occurs:

IP Address:    192.168.1.100     = 11000000.10101000.00000001.01100100
Subnet Mask:   255.255.255.0     = 11111111.11111111.11111111.00000000
                                   |------- Network -------| |--Host--|

Network ID:    192.168.1.0       (IP AND Mask)
Broadcast:     192.168.1.255     (IP OR NOT Mask)
Host Range:    192.168.1.1 - 192.168.1.254

CIDR Notation:

Modern notation uses CIDR (Classless Inter-Domain Routing) prefix notation:

• 192.168.1.0/24 means the first 24 bits are the network portion
• Equivalent to subnet mask 255.255.255.0
• The /24 is called the "prefix length"

IPv4 Address Types:

IPv6 was designed to solve IPv4's address exhaustion with a massively larger address space, while incorporating lessons learned from decades of IPv4 operation.

IPv6 Address Structure:

Scale of IPv6 Address Space:

To appreciate how large 2^128 is:

• If you assigned 1 billion addresses every second since the Big Bang, you'd have used less than 0.00000001% of IPv6 addresses.
• There are enough addresses to assign 100 addresses to every atom on Earth's surface.
• Standard practice allocates a /64 subnet (2^64 = 18 quintillion addresses) to every end-user LAN.

IPv6 Address Abbreviation Rules:

1. Leading zeros in each group can be omitted: 2001:0db8 → 2001:db8
2. One sequence of all-zero groups can be replaced with ::: 2001:db8:0:0:0:0:0:1 → 2001:db8::1
3. The :: can only appear once (to avoid ambiguity)

IPv6 Address Types:

A fundamental advantage of logical addressing is that addresses can be assigned dynamically by software. This section explores the mechanisms that assign IP addresses to hosts.

Static Assignment:

The simplest method—an administrator manually configures each device:

# Linux example
ip addr add 192.168.1.100/24 dev eth0
ip route add default via 192.168.1.1

# Windows example (in Network Adapter Settings)
IP Address: 192.168.1.100
Subnet Mask: 255.255.255.0
Default Gateway: 192.168.1.1
DNS Server: 8.8.8.8

When to use static assignment:

• Servers that need stable, known addresses
• Network infrastructure devices (routers, switches)
• Printers, IoT devices in managed environments
• Systems where DHCP might not be available

Dynamic Host Configuration Protocol (DHCP):

DHCP automates address assignment through a four-step process:

DHCP Provides:

• IP address
• Subnet mask
• Default gateway
• DNS server addresses
• Lease duration (address is borrowed, not permanently assigned)

IPv6 Address Configuration:

IPv6 offers additional configuration methods:

1. SLAAC (Stateless Address Autoconfiguration): Hosts generate their own addresses using:

   • Network prefix learned from router advertisements
   • Interface ID derived from MAC address (EUI-64) or randomly generated (privacy extensions)
   • No DHCP server required for basic connectivity

2. DHCPv6 (Stateful): Similar to DHCPv4, server assigns complete addresses

3. DHCPv6 (Stateless): SLAAC for address, DHCPv6 only for additional info (DNS servers, etc.)

Link-Local Addressing:

When no configuration is available:

• IPv4: APIPA assigns addresses in 169.254.0.0/16
• IPv6: Link-local addresses (fe80::/10) are always automatically configured

These allow local communication even without infrastructure.

Logical addresses enable global routing, but actual data transmission occurs over physical links using physical addresses. The network layer must translate between these two addressing schemes—a process called address resolution.

The Resolution Challenge:

When Host A (192.168.1.10) wants to send a packet to Host B (192.168.1.20):

1. A knows B's IP address
2. A needs B's MAC address to construct the Layer 2 frame
3. There's no formula to derive MAC from IP—they're independent
4. Solution: Ask on the network

Address Resolution Protocol (ARP) for IPv4:

ARP uses a simple request/reply mechanism:

ARP Cache:

Resolving addresses for every packet would be prohibitively slow. Hosts maintain an ARP cache (or ARP table) that stores recent mappings:

$ arp -a
? (192.168.1.1) at d4:e2:5a:1f:3c:88 [ether] on eth0
? (192.168.1.20) at bb:bb:bb:bb:bb:bb [ether] on eth0

Cached entries have a timeout (typically 20 minutes to 4 hours) and are refreshed when needed.

Neighbor Discovery Protocol (NDP) for IPv6:

IPv6 replaces ARP with ICMPv6 Neighbor Discovery:

• Neighbor Solicitation (NS): "What's your link-layer address?" (sent to solicited-node multicast, not broadcast)
• Neighbor Advertisement (NA): "Here's my link-layer address"

NDP advantages over ARP:

• Uses multicast instead of broadcast (more efficient)
• Integrated with ICMPv6 (one less protocol)
• Supports address conflict detection
• Enables router discovery and prefix learning

The true power of logical addressing becomes apparent when we examine how it enables efficient routing. The hierarchical structure of IP addresses is specifically designed to support scalable routing across the global Internet.

Prefix-Based Routing:

Routers don't store entries for every individual host—they match destination addresses against network prefixes:

Routing Table Entry:
  Destination: 192.168.1.0/24
  Next Hop: 10.0.0.2
  Interface: eth1

Meaning: "Any address matching 192.168.1.x should be sent to 10.0.0.2 via eth1"

This single entry covers 254 possible hosts. A /16 prefix covers 65,534 hosts. Aggregation continues upward:

Route Aggregation Example:

Longest Prefix Matching:

When a packet arrives, the router may have multiple matching routes. It selects the most specific match (longest prefix):

Routing Table:
  10.0.0.0/8      → Next hop A (matches 16.7M addresses)
  10.45.0.0/16    → Next hop B (matches 65K addresses)
  10.45.99.0/24   → Next hop C (matches 254 addresses)
  0.0.0.0/0       → Default gateway (matches everything)

Destination: 10.45.99.50
  Matches 10.0.0.0/8 (/8 prefix)
  Matches 10.45.0.0/16 (/16 prefix)
  Matches 10.45.99.0/24 (/24 prefix) ← SELECTED (longest match)
  Matches 0.0.0.0/0 (/0 prefix)

Packet forwarded to Next hop C

Why Hierarchy Matters:

1. Scalability: The global Internet routing table has ~900,000 prefixes, not ~50 billion hosts. This is only possible because addresses aggregate hierarchically.

2. Stability: Changes in individual host addresses don't require global routing updates. Only prefix-level changes propagate.

3. Locality: Traffic destined for the same organization follows the same path until it reaches that organization's network.

4. Policy Enforcement: Network policies can be applied at prefix boundaries (entire organizations, regions, countries).

Logical addressing is the foundation upon which all network layer functions are built. Without hierarchical, assignable, location-aware addresses, global routing would be impossible. Let's consolidate the key insights:

What's Next:

Logical addressing tells us where packets should go; routing tells us how to get them there. The next page explores routing—the mechanisms and algorithms that routers use to discover paths through the network and build the forwarding tables that direct packets hop-by-hop toward their destinations.