Subnets and Masks

Consider an office of 1000 employees on one network. Every broadcast is visible to all, security is nonexistent, performance is terrible. Subnets solve this problem - like walls between departments.

• **VLAN + subnets:** Isolating departments - accounting doesn't see development
• **IP conservation:** ISPs divide a block into small subnets for clients
• **Security:** A DMZ in a separate subnet protects the internal network

Предварительные знания

• IP Addresses: the postal codes of the internet

What is a subnet mask

A **subnet mask** is a 32-bit number that divides an IP address into two parts: the **network** part and the **host** part. The mask bits indicate where the network address ends and the device address begins.

The mask works like a stencil: ones (1) are network bits, zeros (0) are host bits. A bitwise **AND** operation between the IP and the mask gives the network address.

**Why?** The mask determines whether two devices are on the same network. If the network parts match - they can communicate directly. If not - a router is needed.

CIDR notation

**CIDR** (Classless Inter-Domain Routing) is a compact way to write a mask. Instead of `255.255.255.0` we write `/24` - the number of one-bits in the mask. This is called the **prefix length**.

**Why CIDR?** The old class-based system (Class A/B/C) was inflexible. CIDR allows dividing IP space into chunks of any size - precisely matching needs.

**Formula:** number of hosts = 2^(32-prefix) - 2. Minus 2, because the network address and broadcast cannot be assigned to devices.

Network and host bits

Understanding the boundary between network and host bits is the key to working with subnets. Let's work through an example with `/26`.

**Rule:** network address - all host bits are 0; broadcast - all host bits are 1. Everything in between is available for devices.

**Common mistake:** thinking that broadcast always ends in .255. In /26, broadcast is .63; in /28, it's .15. It depends on the subnet size!

Practical subnet calculation

A typical task: "Divide 192.168.1.0/24 into 4 equal subnets". The goal is to determine the new mask and address ranges.

**Tip:** Subnet size is always a power of two: 4, 8, 16, 32, 64, 128, 256. Subnet boundaries are multiples of their size.

Key ideas

• **Mask** divides an IP into network and host parts (AND with IP = network address)
• **CIDR /N** - the number of one-bits in the mask
• **Hosts = 2^(32-N) - 2** (minus network and broadcast)
• **Subnetting:** borrow bits from hosts to get more, smaller networks

Related topics

Subnets are the foundation of network architecture:

• Private IP addresses — Which ranges to use inside a network
• Routing — How a router determines a path using a mask
• VLSM — Variable-length subnets for optimization

Вопросы для размышления

• Why can't the broadcast address be assigned to a device?
• How to tell if two IP addresses are on the same subnet without a calculator?
• Why use /30 for point-to-point connections?

Связанные уроки

• alg-35-bit-manipulation