IP Addresses and Addressing

Chapter 3 separated roles, hosts, and hostnames. It also left a deliberate gap: a hostname is a name used to refer to a host, but hosts that participate in IP networking also use addresses.

People often fill that gap with one wrong sentence: “the IP address of the computer.” That sentence is common. It is also too strong. This chapter opens the address plane carefully—enough for operators to stop confusing names, hosts, and addresses—without becoming a networking textbook.

Why This Matters

Email systems communicate across IP networks. Operators repeatedly see both hostnames and IP addresses in logs, configuration, and connection details. If those identifiers collapse into one idea—“the server”—diagnosis starts in the wrong place.

Later Handbook chapters will talk about connections, paths, DNS associations, and operational IP practices. Those chapters become clearer if this one installs one durable habit first:

Names and Addresses Solve Different Problems

Internet Protocol language has long distinguished names, addresses, and routes: a name indicates what we seek; an address indicates where it is; a route indicates how to get there (RFC 791).

Chapter 3 already used the name side. Domain-name design goals likewise include that names should not be required to contain network identifiers, addresses, or routes (RFC 1034).

So keep the separation clean:

A hostname is a name used to refer to a host.

An IP address is a network-layer address used in IP communication.

They can be associated. They are not the same kind of identifier. The association is not required to be permanent, singular, or one-to-one.

What Is an IP Address?

An IP address is an address used by the Internet Protocol: a network-layer identifier that participates in addressing traffic on an IP network (RFC 791; RFC 8200).

It is not a product name. It is not a role. It is not a hostname written in a different alphabet.

Two address families matter at Foundations depth:

IPv4 — 32-bit addresses, commonly written as four decimal numbers separated by dots (RFC 791).

IPv6 — 128-bit addresses, conventionally written with hexadecimal groups separated by colons (RFC 4291; RFC 8200).

IPv6 is not “IPv4 with more digits.” It is another IP address family with a different size and addressing architecture. For this chapter, you only need to recognize both families and understand that a host environment may use either—or both.

What Does an Address Actually Identify?

The beginner trap is absolute:

Useful as slang. Misleading as architecture.

A host participates in an IP network through one or more network attachments (often called interfaces); IP addresses are associated with those attachments, not with the machine as a permanent whole (RFC 4291; RFC 1122).

Empire’s Foundations framing is therefore:

That is why one host can use more than one address—and why two different addresses do not automatically mean two different machines.

IPv4 and IPv6 at Map Depth

IPv4 remains widely visible in mail infrastructure. A typical written form looks like `192.0.2.10`. That example uses documentation address space reserved for examples, not a random public address (RFC 5737).

IPv6 uses a larger address size and a different written convention, for example `2001:DB8::1` from the documentation prefix reserved for examples (RFC 3849; RFC 4291). Text forms can be shortened in practice; Foundations does not need compression drills.

A host may use IPv4, IPv6, or both. When both are present, operators often call that coexistence dual-stack. Keep the idea light: two address families can be available in the same environment. Transition mechanisms stay later.

One Host, Multiple Addresses

Because addresses attach to network attachments—and because a host may have more than one attachment or more than one address on an attachment—one host can use multiple IP addresses (RFC 1122; RFC 4291).

HOST
├── hostname(s)
└── IP address(es)
    ├── IPv4 address
    ├── another IPv4 address (possible)
    └── IPv6 address (possible)

Read this as vocabulary hygiene, not a floor plan. Names and addresses need not map one-to-one. They can change. Seeing two addresses does not prove you are looking at two physical machines.

A single network attachment may also carry multiple addresses (RFC 4291). That reinforces the same lesson without requiring configuration depth.

Addresses Are Assigned and Can Change

IP addresses are assigned. Assignments can change over time.

Some addresses remain stable for long periods by operational choice. Others do not. Foundations only needs the anti-myth:

How assignment happens—leases, automation, cloud renumbering—belongs later. The diagnostic habit belongs now.

Public and Private IPv4

Some IPv4 address blocks are reserved for private internets. The reserved private ranges are `10.0.0.0`–`10.255.255.255`, `172.16.0.0`–`172.31.255.255`, and `192.168.0.0`–`192.168.255.255` (RFC 1918).

Private addresses are for use inside private networks. They are not globally unique identifiers for the public Internet (RFC 1918).

Two moral shortcuts fail:

Private does not mean secure.

Public does not mean unsafe.

“Public” and “private” here are about global uniqueness and ordinary Internet reachability expectations—not about trustworthiness.

Do not treat address translation as the definition of private addressing; translation mechanics are later. Also do not casually project “private IPv4” language onto IPv6; IPv6 has its own local-address concepts that this chapter leaves alone.

Teaching Model: Role, Host, Hostname, Address

In plain language:

Role — what function is being performed (Chapter 2).

Host — the networked system where software can run (Chapter 3).

Hostname — a name used to refer to a host (Chapter 3).

IP address — a network-layer address associated with a host’s network attachment.

These abstractions relate. They are not synonyms. Do not force a one-to-one ladder that pretends every hostname always maps to exactly one permanent address.

Where Names Meet Addresses

DNS can associate domain names with resource information, including addresses. Chapter 3 introduced that naming infrastructure only as orientation. This chapter adds only one continuity sentence:

How those associations are published and looked up—record types and tooling—belongs to later DNS chapters. Do not treat DNS as the system that “gives a machine its IP address.” Assignment and naming association are different jobs.

Why Email Operators Care

Internet Mail moves across independently operated systems on IP networks. Operators continually encounter hostnames and IP addresses while reading infrastructure descriptions and connection evidence.

The Foundations need is identifier hygiene:

Am I looking at a role?

Am I looking at a host?

Am I looking at a hostname?

Am I looking at an IP address?

Confusing those categories invents false certainty—“this must be a different server,” “this must be permanent,” “this must be the application,” “this must be the domain’s dedicated identity”—before evidence supports it.

Deliverability strategy, reputation, and reverse-DNS practice come later. They become safer to learn after this model is stable.

Operator Perspective

Chapter 2 asked:

Chapter 3 asked:

Chapter 4 adds:

Role first. Host and hostname second. Address third. Product labels and dashboard strings after that.

Common Misconceptions

Key Takeaways