Network Layer Addressing

Every device connected to the Internet possesses a unique identifier that enables it to send and receive data from any other device worldwide. This identifier is not the hardware address burned into your network card—it's something more abstract, more flexible, and infinitely more powerful: a logical address.

Understanding logical addresses is fundamental to comprehending how the Internet works. Without them, the seamless global communication we take for granted—streaming video from servers in another continent, sending emails across the world, accessing cloud services hosted anywhere—would be impossible.

This page explores the concept of logical addressing in depth, examining why physical addresses alone are insufficient for internetworking, how logical addresses solve these limitations, and the fundamental principles that govern address design in modern networks.

Before we can appreciate the elegance of logical addressing, we must understand why physical addresses—though essential at the data link layer—are fundamentally inadequate for internetworking.

Physical Addresses (MAC Addresses):

Every network interface card (NIC) is manufactured with a unique identifier called a MAC (Media Access Control) address. This 48-bit address (typically displayed as six hexadecimal pairs like 00:1A:2B:3C:4D:5E) is burned into the hardware during manufacturing.

MAC addresses excel at local network communication. Within a single LAN segment, devices use MAC addresses to identify each other, with Ethernet frames carrying source and destination MAC addresses for direct delivery.

However, MAC addresses present critical limitations when we attempt to scale beyond a single network:

The Thought Experiment:

Consider the implications of routing with MAC addresses alone:

1. Global Table Size: With approximately 20 billion connected devices, every router would need routing entries for each device
2. No Path Aggregation: A router in New York couldn't know that all devices starting with certain MAC prefixes are in Tokyo—manufacturers ship NICs worldwide
3. Constant Flux: Devices move, networks change, but MAC addresses stay fixed—routing tables would be perpetually outdated
4. Lookup Complexity: Even with perfect information, searching 20 billion entries for every packet would consume impossible amounts of time and memory

These limitations aren't engineering challenges to be overcome—they're fundamental architectural mismatches between physical addressing and internetwork routing requirements. The solution demands a different kind of address entirely.

A logical address (also called a network address or IP address) is a software-assigned identifier that provides location within a network topology rather than hardware identity. This distinction is profound and enables everything we associate with modern internetworking.

Definition and Characteristics:

Logical addresses are:

• Software-assigned: Configured by network administrators or assigned dynamically, independent of hardware
• Location-specific: Reflect where a device is connected within the network topology
• Hierarchically structured: Designed to enable aggregation, summarization, and efficient routing
• Mutable: Can be changed when devices move between networks
• Protocol-specific: Defined by the network layer protocol (IP in TCP/IP networks)

The Postal Analogy Revisited:

If MAC addresses are like unique serial numbers stamped on each house (useful for inventory, useless for delivery), then IP addresses are like proper postal addresses:

• Country → Network class/organization allocation
• City → Major network identifier
• Street → Subnet identifier
• House number → Host identifier

Just as a letter travels from country to city to street to house, an IP packet routes from global networks to regional networks to local subnets to individual hosts. Each level of the hierarchy narrows the search space dramatically.

Key Insight: Logical addresses answer the question "Where is this device located in the network topology?" while physical addresses answer "Which piece of hardware is this?" Both questions matter, but only the first enables routing.

The Internet Protocol (IP) defines the logical addressing scheme used throughout the global Internet and most private networks. Understanding IP addresses is essential for any network engineer, system administrator, or software developer working with networked systems.

IPv4 Address Structure:

IPv4 addresses are 32-bit binary numbers, typically displayed in dotted-decimal notation for human readability:

Binary:  11000000.10101000.00000001.01100100
Decimal: 192.168.1.100

Each of the four octets represents 8 bits, ranging from 0 to 255. This gives IPv4 a theoretical maximum of 2³² = 4,294,967,296 unique addresses.

IPv6 Address Structure:

Recognizing IPv4's address space limitations, IPv6 expanded to 128-bit addresses, displayed in hexadecimal notation:

2001:0db8:85a3:0000:0000:8a2e:0370:7334

This provides 2¹²⁸ ≈ 3.4 × 10³⁸ addresses—enough to assign billions of addresses to every grain of sand on Earth.

The Network-Host Division:

Every IP address is divided into two logical components:

1. Network Portion: Identifies the network the device belongs to
2. Host Portion: Identifies the specific device within that network

The boundary between network and host portions is determined by the subnet mask (IPv4) or prefix length (both versions):

IP Address:    192.168.1.100
Subnet Mask:   255.255.255.0  (or /24)
                
Network Part:  192.168.1.___  (first 24 bits)
Host Part:     ___.___.___.100 (last 8 bits)

This division is the architectural foundation of hierarchical addressing—routers only need to know how to reach networks, not individual hosts.

To fully appreciate logical addressing, let's trace how it enables packet delivery across the Internet—something physical addresses alone could never accomplish.

Scenario: Cross-Country Web Request

You're in New York, accessing a website hosted in San Francisco:

• Your laptop: IP 203.0.113.50 on network 203.0.113.0/24
• Web server: IP 198.51.100.200 on network 198.51.100.0/24

The Routing Journey:

Step-by-Step Analysis:

Step 1: Local Decision Your laptop compares the destination IP (198.51.100.200) against its own network (203.0.113.0/24). Since the destination is on a different network, the packet must be sent to the default gateway.

Step 2: Gateway Routing
Your ISP's router examines its routing table. It doesn't know about 198.51.100.200 specifically, but it has an entry for 198.0.0.0/8 pointing toward the Internet backbone.

Step 3: Progressive Refinement
As the packet traverses core routers, each has increasingly specific routing information. One router knows 198.51.0.0/16 is in California. Another knows 198.51.100.0/24 is reachable through a specific ISP.

Step 4: Final Delivery
The destination network's router has entries for all hosts on 198.51.100.0/24. It delivers the packet directly to 198.51.100.200.

Key Observations:

• No router needed to know about your specific laptop or the server's MAC address
• Each router made decisions based only on the network portion of the destination IP
• Hierarchical aggregation meant core routers only needed a handful of routes representing millions of hosts
• The packet traversed heterogeneous networks (Ethernet, fiber, MPLS) seamlessly—because logical addresses transcend physical media

A common question from students is: "If logical addresses are so superior, why do we still need physical addresses at all?"

The answer reveals a fundamental principle of layered network design: each layer serves a different purpose, and addresses at each layer solve different problems.

Why Both Addresses Are Necessary:

The Handoff Process:

Consider what happens when a packet arrives at the final router before its destination:

1. Router receives a packet destined for IP 192.168.1.100
2. Router determines 192.168.1.100 is on a directly connected network
3. Router needs to deliver the packet over Ethernet—but Ethernet uses MAC addresses
4. Router consults its ARP cache (or sends an ARP request) to find the MAC address corresponding to 192.168.1.100
5. Router encapsulates the IP packet in an Ethernet frame addressed to the host's MAC address
6. The Ethernet switch delivers the frame to the correct physical port

This handoff between logical and physical addressing happens at every hop, with MAC addresses changing at each router while the IP addresses remain constant end-to-end.

The Encapsulation Model:

┌────────────────────────────────────────────────────────┐
│ Ethernet Frame (Layer 2)                               │
│ ┌────────────┬──────────────────────────────────────┐ │
│ │ MAC Header │ IP Packet (Layer 3)                  │ │
│ │ Src: R1    │ ┌────────────┬───────────────────┐   │ │
│ │ Dst: Host  │ │ IP Header  │ Transport + Data  │   │ │
│ │            │ │ Src: A     │                   │   │ │
│ │            │ │ Dst: B     │                   │   │ │
│ └────────────┴─┴────────────┴───────────────────┘   │ │
└────────────────────────────────────────────────────────┘

The IP addresses (A → B) remain unchanged across the entire journey. The MAC addresses change at every hop, reflecting the current physical segment's source and destination.

Effective logical addressing requires careful design. IP addresses must satisfy several critical properties to enable reliable, scalable internetworking:

To solidify understanding, let's develop robust mental models for thinking about logical and physical addresses.

Mental Model 1: Identity vs Location

Think of it like a person's identity versus their current whereabouts:

• MAC Address = Your fingerprint (permanent, unique, identifies you)
• IP Address = Your current postal address (changes when you move, enables delivery)

Your fingerprint is always the same, but when you travel from home to a hotel, your package delivery address changes. Both identifiers are useful for different purposes.

Mental Model 2: Local vs Global Names

Consider naming conventions in organizations:

• MAC Address = First name (works within a room: "Hey, John!")
• IP Address = Full corporate directory path (works company-wide: "John Smith, Engineering, Building C, Floor 3")

In a small meeting room, first names suffice. Across a 10,000-person company, structured naming becomes essential.

Mental Model 3: Train Journey Tickets

Imagine traveling by train across multiple countries:

• IP Address = Your final destination ticket (Paris → Berlin)
• MAC Address = Your seat assignment for each train segment (changes at each transfer)

The conductor in each train car only cares about seat assignments (local delivery), but the overall routing system uses your destination (global routing).

We've explored the foundational concept of logical addressing—the mechanism that makes internetworking possible. Let's consolidate the essential insights:

What's Next:

With the concept of logical addressing established, the next page explores hierarchical addressing—the design principle that gives IP addresses their power. You'll learn how the division of addresses into network and host portions, combined with subnetting and CIDR, enables the Internet to scale from small home networks to global infrastructure.