Computer NetworksCIDR

CIDR - Classless Inter-Domain Routing

LevelIntermediate

Duration60 mins

TopicCIDR

1 / 5

Classless Inter-Domain Routing

The Birth of a Classless Internet

In the early 1990s, the Internet faced an existential crisis. The classful addressing system—with its rigid Class A, B, and C boundaries—was hemorrhaging IP addresses at an alarming rate. Organizations requesting just 500 addresses were allocated an entire Class B network of 65,536 addresses, wasting over 99% of the allocation. Meanwhile, the global routing tables were exploding in size, threatening to overwhelm router memory and processing capabilities.

The solution that emerged wasn't just an incremental improvement—it was a paradigm shift in how we think about IP addressing. That solution was Classless Inter-Domain Routing (CIDR).

What You Will Learn

By the end of this page, you will understand why CIDR was developed, how it fundamentally differs from classful addressing, the problems it solves, and why it remains the foundation of modern Internet addressing. You'll grasp the historical context that made CIDR necessary and the elegant mathematical principles that make it work.

The Classful Crisis

To understand CIDR, we must first appreciate the magnitude of the problem it solved. The original IPv4 addressing scheme, designed in 1981, divided the 32-bit address space into five rigid classes:

The Fundamental Problem: Granularity Mismatch

Classful addressing offered only three usable network sizes for organizations:

Class A: 16,777,214 hosts (for massive networks)
Class B: 65,534 hosts (for medium networks)
Class C: 254 hosts (for small networks)

But real-world organizations don't fit into neat categories. What about a company needing 1,000 addresses? Or 10,000? The classful system forced impossible choices.

The Classful Addressing Waste Problem
Hosts Needed	Class Assigned	Addresses Allocated	Addresses Wasted	Efficiency
500 hosts	Class B	65,534	65,034	0.76%
2,000 hosts	Class B	65,534	63,534	3.05%
10,000 hosts	Class B	65,534	55,534	15.26%
300 hosts	2× Class C	508	208	59.06%
1,000 hosts	4× Class C	1,016	16	98.4%

The Two-Headed Monster

By the early 1990s, the Internet faced two simultaneous crises:

Address Depletion: Class B addresses were being exhausted rapidly. Organizations that needed more than 254 hosts but fewer than 65,534 were forced to request Class B allocations, wasting enormous chunks of the address space.
Routing Table Explosion: Attempts to conserve Class B addresses by assigning multiple Class C networks created a different problem—routing table explosion. An organization needing 2,000 addresses might receive 8 Class C networks, adding 8 separate entries to the global routing table instead of 1.

The math was brutal: if the Internet continued on its trajectory, routing tables would exceed router memory by 1995, and Class B addresses would be completely exhausted by 1994.

Historical Context: The 1994 Deadline

In 1992, the IETF projected that at the current allocation rate, Class B addresses would be completely exhausted by March 1994—just 18 months away. This wasn't theoretical; it was an imminent catastrophe that would have frozen Internet growth. CIDR was developed and deployed under extreme time pressure.

The Classless Revolution

CIDR, defined in RFC 1518 and RFC 1519 (September 1993), represented a fundamental reconceptualization of IP addressing. Instead of the rigid class-based hierarchy, CIDR introduced a simple but profound idea: the network/host boundary can be placed at any bit position in the 32-bit address.

The Core Insight

In classful addressing, the first few bits of an IP address determined its class and, by extension, its network/host boundary:

Class A: /8 boundary (first 8 bits = network)
Class B: /16 boundary (first 16 bits = network)
Class C: /24 boundary (first 24 bits = network)

CIDR eliminated these artificial constraints entirely. Instead, the network portion of an address could be any number of bits from 1 to 32, specified explicitly with a prefix length.

CIDR vs Classful: The Conceptual Shift
┌─────────────────────────────────────────────────────────────────┐
│ CLASSFUL ADDRESSING: Implicit, Rigid Boundaries                 │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  IP Address: 192.168.1.0                                        │
│                                                                 │
│  First bits reveal class:                                       │
│  110xxxxx = Class C                                             │
│                                                                 │
│  Therefore: Network = 192.168.1.0 (first 24 bits)               │
│             Hosts   = .0 to .255 (last 8 bits)                  │
│             Size    = 256 addresses (fixed!)                    │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘
 
┌─────────────────────────────────────────────────────────────────┐
│ CIDR ADDRESSING: Explicit, Flexible Boundaries                  │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  192.168.0.0/22  →  Network = first 22 bits                     │
│                     Hosts   = last 10 bits                      │
│                     Size    = 2^10 = 1,024 addresses            │
│                                                                 │
│  192.168.0.0/23  →  Network = first 23 bits                     │
│                     Hosts   = last 9 bits                       │
│                     Size    = 2^9 = 512 addresses               │
│                                                                 │
│  192.168.0.0/25  →  Network = first 25 bits                     │
│                     Hosts   = last 7 bits                       │
│                     Size    = 2^7 = 128 addresses               │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

The Power of Flexibility

This seemingly simple change had profound implications:

Right-Sized Allocations: An organization needing 1,000 addresses could now receive exactly a /22 (1,024 addresses) instead of wasting 64,000+ addresses on a Class B.
Arbitrary Subdivision: Large allocations could be subdivided at any granularity. A /16 (65,536 addresses) could be split into 256 /24s, 64 /22s, or any other combination.
Address Recovery: Unused portions of existing allocations could be reclaimed and reallocated without breaking the addressing structure.

But the true genius of CIDR wasn't just in flexible allocation—it was in how this flexibility enabled address aggregation, which we'll explore in subsequent pages.

The Name Matters

CIDR is called 'Classless Inter-Domain Routing' because it operates between autonomous systems (domains) on the Internet, enabling aggregation at network boundaries. The 'classless' part emphasizes the complete rejection of the A/B/C/D/E classification scheme. Every address is treated the same way—as a base address plus a prefix length.

Mathematical Foundation of CIDR

CIDR's power derives from the mathematical properties of binary numbers and prefix-based addressing. Understanding these foundations is essential for working with CIDR in practice.

The Prefix Length

A CIDR prefix length (the number after the /) specifies how many bits of the address are fixed (the network portion). The remaining bits can vary (the host portion).

/n means the first n bits are the network identifier
The remaining 32 - n bits are available for host addresses
The total addresses in a CIDR block = 2^(32-n)

CIDR Prefix Lengths and Address Counts
Prefix	Network Bits	Host Bits	Addresses	Subnet Mask	Classful Equivalent
/8	8	24	16,777,216	255.0.0.0	1 Class A
/16	16	16	65,536	255.255.0.0	1 Class B
/20	20	12	4,096	255.255.240.0	16 Class C
/22	22	10	1,024	255.255.252.0	4 Class C
/24	24	8	256	255.255.255.0	1 Class C
/26	26	6	64	255.255.255.192	1/4 Class C
/28	28	4	16	255.255.255.240	1/16 Class C
/30	30	2	4	255.255.255.252	Point-to-point
/32	32	0	1	255.255.255.255	Single host

Block Alignment: The Critical Constraint

Not every IP address can serve as the start of any CIDR block. A CIDR block must be properly aligned—meaning the starting address must be divisible by the block size.

For example:

192.168.0.0/24 is valid (192.168.0.0 is divisible by 256)
192.168.1.0/24 is valid (192.168.1.0 is divisible by 256)
192.168.0.128/24 is invalid (192.168.0.128 is not divisible by 256)

The alignment rule ensures that CIDR blocks don't overlap in unintended ways and can be aggregated cleanly.

Block Alignment Examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
# Understanding CIDR Block Alignment
# A block with prefix /n must start at an address divisible by 2^(32-n)
 
Example: /22 block (1,024 addresses)
Block size = 2^(32-22) = 2^10 = 1,024
 
Valid /22 blocks:
192.168.0.0/22   →  192.168.0.0 = 3232235520 (divisible by 1024 ✓)
192.168.4.0/22   →  192.168.4.0 = 3232236544 (divisible by 1024 ✓)
192.168.8.0/22   →  192.168.8.0 = 3232237568 (divisible by 1024 ✓)
 
Invalid /22 blocks:
192.168.1.0/22   →  192.168.1.0 = 3232235776 (not divisible by 1024 ✗)
192.168.2.0/22   →  192.168.2.0 = 3232236032 (not divisible by 1024 ✗)
192.168.3.0/22   →  192.168.3.0 = 3232236288 (not divisible by 1024 ✗)
 
Binary representation of valid /22 boundaries:
192.168.0.0   = 11000000.10101000.000000|00.00000000
192.168.4.0   = 11000000.10101000.000001|00.00000000
192.168.8.0   = 11000000.10101000.000010|00.00000000
                                        ^
                                The /22 boundary (last 10 bits = host portion)

The Binary Perspective

In binary, a properly aligned CIDR block always has zeros in all the host bits of its starting address. This is why 192.168.4.0/22 is valid (binary ends in 10 zeros for host portion) but 192.168.1.0/22 is not (the .1 means some host bits are set in the base address).

CIDR vs Classful: A Deep Comparison

Understanding the fundamental differences between classful and classless addressing clarifies why CIDR represented such a revolutionary change in Internet architecture.

Classful Addressing

•Network mask is implicit (determined by first bits)
•Only 3 network sizes: /8, /16, /24
•Wastes addresses in most allocations
•Every network = separate routing entry
•No hierarchical aggregation possible
•First octet determines network class
•Address space divided into 5 rigid classes
•Subnetting hidden from external routers

CIDR (Classless)

•Network mask is explicit (/n notation)
•Any network size from /1 to /32
•Allocations match actual needs precisely
•Multiple networks aggregated into one entry
•Hierarchical routing through aggregation
•First octet has no special meaning
•Address space treated as a continuum
•All prefix information propagated globally

The Routing Information Problem

Consider how routing changes between the two systems:

Classful Scenario: An ISP needs to advertise connectivity to 16 customers, each with a Class C network.

Routing table entries: 16 separate routes
Each route: 192.168.0.0, 192.168.1.0, ..., 192.168.15.0 (all /24)
No reduction possible

CIDR Scenario: The same 16 networks, if contiguous, become:

Routing table entries: 1 aggregated route
Single route: 192.168.0.0/20 (covers all 16 /24 networks)
93.75% reduction in routing table size

This aggregation capability is what saved the Internet from routing table collapse.

The CIDR Success Story

Within two years of CIDR deployment (1993-1995), the exponential growth of Internet routing tables slowed dramatically. What would have reached 100,000+ entries was held under 40,000 through aggressive aggregation. CIDR bought the Internet enough time for IPv6 development and continues to be the foundation of address management today.

How CIDR Address Allocation Works

CIDR transformed not just the technical format of addresses, but the entire ecosystem of address allocation. Understanding this hierarchy is essential for network engineers.

The Hierarchical Allocation Model

CIDR enabled a clean hierarchical distribution of address space:

IANA (Internet Assigned Numbers Authority)
- Controls the entire IPv4 address space
- Allocates large blocks (/8 or larger) to Regional Internet Registries
RIRs (Regional Internet Registries)
- Five organizations covering global regions (ARIN, RIPE, APNIC, LACNIC, AFRINIC)
- Receive large blocks from IANA, allocate to ISPs and large organizations
- Typical allocations: /12 to /19
ISPs (Internet Service Providers)
- Receive blocks from RIRs
- Sub-allocate to customers
- Aggregate customer routes into single advertisements
- Typical allocations to customers: /24 to /29
End Organizations
- Receive address blocks from ISPs
- May further subnet internally
- Do not advertise to the global Internet directly

Converting Mermaid diagram...

Provider-Aggregatable (PA) vs Provider-Independent (PI) Addresses

CIDR creates two categories of address allocations:

Provider-Aggregatable (PA):

Addresses allocated from an ISP's block
Can be aggregated by the ISP (reducing global routing table)
If customer changes ISPs, addresses must be returned
Most common for typical organizations

Provider-Independent (PI):

Addresses allocated directly from an RIR
Cannot be aggregated (appear as separate routes)
Customer keeps addresses when changing ISPs
Reserved for organizations with strong justification
Contributes to routing table growth

The Aggregation Dependency

CIDR's effectiveness depends on addresses being allocated contiguously and hierarchically. When addresses are fragmented (due to PI allocations, mergers, or historical assignments), aggregation becomes impossible, and routing table growth accelerates. This is why RIRs carefully manage address allocation policies.

CIDR in Modern Networks

Three decades after its introduction, CIDR remains the foundational addressing scheme for the Internet and virtually all IP networks. Its principles have been extended and adapted to modern networking contexts.

Cloud Computing and CIDR

Every major cloud provider uses CIDR extensively:

VPC/VNet design: Cloud virtual networks are defined using CIDR blocks
Subnet creation: VPCs are subdivided using smaller CIDR blocks
Security groups: Firewall rules reference CIDR blocks for source/destination
Route tables: All entries use CIDR notation

CIDR in Cloud Architecture (AWS Example)

Terraform

# AWS VPC Design Using CIDR Blocks
# Demonstrates hierarchical CIDR subdivision
 
resource "aws_vpc" "main" {
  cidr_block = "10.0.0.0/16"  # 65,536 addresses for entire VPC
  
  tags = {
    Name = "Production-VPC"
  }
}
 
# Public subnets: 10.0.0.0/20 - 10.0.15.255 (4,096 addresses each)
resource "aws_subnet" "public_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.0.0/20"   # 10.0.0.0 - 10.0.15.255
  availability_zone = "us-east-1a"
}
 
resource "aws_subnet" "public_az2" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.16.0/20"  # 10.0.16.0 - 10.0.31.255
  availability_zone = "us-east-1b"
}
 
# Private subnets: larger blocks for application servers
resource "aws_subnet" "private_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.128.0/18" # 10.0.128.0 - 10.0.191.255 (16K addresses)
  availability_zone = "us-east-1a"
}
 
resource "aws_subnet" "private_az2" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.192.0/18" # 10.0.192.0 - 10.0.255.255 (16K addresses)
  availability_zone = "us-east-1b"
}
 
# Database subnets: smaller, isolated blocks
resource "aws_subnet" "database_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.32.0/24"  # 10.0.32.0 - 10.0.32.255 (256 addresses)
  availability_zone = "us-east-1a"
}
 
# Security group using CIDR for access control
resource "aws_security_group_rule" "allow_internal" {
  type        = "ingress"
  from_port   = 0
  to_port     = 65535
  protocol    = "tcp"
  cidr_blocks = ["10.0.0.0/16"]  # Allow all traffic within VPC
  security_group_id = aws_security_group.internal.id
}
 
resource "aws_security_group_rule" "allow_partners" {
  type        = "ingress"
  from_port   = 443
  to_port     = 443
  protocol    = "tcp"
  cidr_blocks = [
    "203.0.113.0/24",    # Partner A network
    "198.51.100.0/25",   # Partner B network
  ]
  security_group_id = aws_security_group.api.id
}

Container Networking and CIDR

Container orchestration platforms like Kubernetes use CIDR extensively:

Pod CIDR: Each node receives a CIDR block for pod IP addresses
Service CIDR: Virtual IPs for services come from a dedicated block
Network Policies: Access control rules use CIDR notation

IPv6 and CIDR

CIDR concepts carry directly to IPv6, which was designed with CIDR from the start:

IPv6 prefix notation: 2001:db8::/32 (standard allocation size)
No classes ever existed in IPv6
Typically, enterprises receive a /48 (280 trillion subnets possible)
End sites often receive a /64 (standard subnet size)

CIDR Notation is Universal

Whether you're configuring a home router, designing an enterprise network, deploying cloud infrastructure, or writing firewall rules, CIDR notation is the universal language. Mastering it pays dividends across every area of networking.

Summary: The Classless Revolution

CIDR represents one of the most significant evolutionary steps in Internet architecture. By eliminating the artificial class boundaries and enabling flexible prefix lengths, CIDR solved the immediate crises of address depletion and routing table explosion while providing a framework that continues to scale today.

Key Takeaways

•CIDR eliminated class boundaries — The network/host division can occur at any bit position, not just at /8, /16, or /24.
•Prefix length is explicit — The /n notation clearly specifies the network portion, removing ambiguity.
•Block sizes follow powers of 2 — A /n block contains 2^(32-n) addresses, enabling predictable calculations.
•Alignment constraints apply — CIDR blocks must start at addresses divisible by their size.
•Hierarchical allocation enables aggregation — Contiguous allocations can be summarized into single routing entries.
•CIDR saved the Internet — Without it, address exhaustion and routing collapse would have occurred in the mid-1990s.

What's Next:

Now that we understand what CIDR is and why it was developed, we'll dive into the practical mechanics. The next page covers prefix notation in depth—how to read, write, and calculate CIDR blocks, with hands-on examples that will make you fluent in /n notation.

Page Complete

You now understand the foundational concepts of CIDR—why it was created, how it differs from classful addressing, and its ongoing importance in modern networking. The next page will build on this foundation with practical prefix notation skills.

1 / 5

Loading learning content...

Computer NetworksCIDR

CIDR - Classless Inter-Domain Routing

LevelIntermediate

Duration60 mins

TopicCIDR

1 / 5

Classless Inter-Domain Routing

The Birth of a Classless Internet

The solution that emerged wasn't just an incremental improvement—it was a paradigm shift in how we think about IP addressing. That solution was Classless Inter-Domain Routing (CIDR).

What You Will Learn

The Classful Crisis

To understand CIDR, we must first appreciate the magnitude of the problem it solved. The original IPv4 addressing scheme, designed in 1981, divided the 32-bit address space into five rigid classes:

The Fundamental Problem: Granularity Mismatch

Classful addressing offered only three usable network sizes for organizations:

Class A: 16,777,214 hosts (for massive networks)
Class B: 65,534 hosts (for medium networks)
Class C: 254 hosts (for small networks)

But real-world organizations don't fit into neat categories. What about a company needing 1,000 addresses? Or 10,000? The classful system forced impossible choices.

The Classful Addressing Waste Problem
Hosts Needed	Class Assigned	Addresses Allocated	Addresses Wasted	Efficiency
500 hosts	Class B	65,534	65,034	0.76%
2,000 hosts	Class B	65,534	63,534	3.05%
10,000 hosts	Class B	65,534	55,534	15.26%
300 hosts	2× Class C	508	208	59.06%
1,000 hosts	4× Class C	1,016	16	98.4%

The Two-Headed Monster

By the early 1990s, the Internet faced two simultaneous crises:

Address Depletion: Class B addresses were being exhausted rapidly. Organizations that needed more than 254 hosts but fewer than 65,534 were forced to request Class B allocations, wasting enormous chunks of the address space.
Routing Table Explosion: Attempts to conserve Class B addresses by assigning multiple Class C networks created a different problem—routing table explosion. An organization needing 2,000 addresses might receive 8 Class C networks, adding 8 separate entries to the global routing table instead of 1.

The math was brutal: if the Internet continued on its trajectory, routing tables would exceed router memory by 1995, and Class B addresses would be completely exhausted by 1994.

Historical Context: The 1994 Deadline

The Classless Revolution

The Core Insight

In classful addressing, the first few bits of an IP address determined its class and, by extension, its network/host boundary:

Class A: /8 boundary (first 8 bits = network)
Class B: /16 boundary (first 16 bits = network)
Class C: /24 boundary (first 24 bits = network)

CIDR eliminated these artificial constraints entirely. Instead, the network portion of an address could be any number of bits from 1 to 32, specified explicitly with a prefix length.

CIDR vs Classful: The Conceptual Shift
┌─────────────────────────────────────────────────────────────────┐
│ CLASSFUL ADDRESSING: Implicit, Rigid Boundaries                 │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  IP Address: 192.168.1.0                                        │
│                                                                 │
│  First bits reveal class:                                       │
│  110xxxxx = Class C                                             │
│                                                                 │
│  Therefore: Network = 192.168.1.0 (first 24 bits)               │
│             Hosts   = .0 to .255 (last 8 bits)                  │
│             Size    = 256 addresses (fixed!)                    │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘
 
┌─────────────────────────────────────────────────────────────────┐
│ CIDR ADDRESSING: Explicit, Flexible Boundaries                  │
├─────────────────────────────────────────────────────────────────┤
│                                                                 │
│  192.168.0.0/22  →  Network = first 22 bits                     │
│                     Hosts   = last 10 bits                      │
│                     Size    = 2^10 = 1,024 addresses            │
│                                                                 │
│  192.168.0.0/23  →  Network = first 23 bits                     │
│                     Hosts   = last 9 bits                       │
│                     Size    = 2^9 = 512 addresses               │
│                                                                 │
│  192.168.0.0/25  →  Network = first 25 bits                     │
│                     Hosts   = last 7 bits                       │
│                     Size    = 2^7 = 128 addresses               │
│                                                                 │
└─────────────────────────────────────────────────────────────────┘

The Power of Flexibility

This seemingly simple change had profound implications:

Right-Sized Allocations: An organization needing 1,000 addresses could now receive exactly a /22 (1,024 addresses) instead of wasting 64,000+ addresses on a Class B.
Arbitrary Subdivision: Large allocations could be subdivided at any granularity. A /16 (65,536 addresses) could be split into 256 /24s, 64 /22s, or any other combination.
Address Recovery: Unused portions of existing allocations could be reclaimed and reallocated without breaking the addressing structure.

But the true genius of CIDR wasn't just in flexible allocation—it was in how this flexibility enabled address aggregation, which we'll explore in subsequent pages.

The Name Matters

Mathematical Foundation of CIDR

CIDR's power derives from the mathematical properties of binary numbers and prefix-based addressing. Understanding these foundations is essential for working with CIDR in practice.

The Prefix Length

A CIDR prefix length (the number after the /) specifies how many bits of the address are fixed (the network portion). The remaining bits can vary (the host portion).

/n means the first n bits are the network identifier
The remaining 32 - n bits are available for host addresses
The total addresses in a CIDR block = 2^(32-n)

CIDR Prefix Lengths and Address Counts
Prefix	Network Bits	Host Bits	Addresses	Subnet Mask	Classful Equivalent
/8	8	24	16,777,216	255.0.0.0	1 Class A
/16	16	16	65,536	255.255.0.0	1 Class B
/20	20	12	4,096	255.255.240.0	16 Class C
/22	22	10	1,024	255.255.252.0	4 Class C
/24	24	8	256	255.255.255.0	1 Class C
/26	26	6	64	255.255.255.192	1/4 Class C
/28	28	4	16	255.255.255.240	1/16 Class C
/30	30	2	4	255.255.255.252	Point-to-point
/32	32	0	1	255.255.255.255	Single host

Block Alignment: The Critical Constraint

Not every IP address can serve as the start of any CIDR block. A CIDR block must be properly aligned—meaning the starting address must be divisible by the block size.

For example:

192.168.0.0/24 is valid (192.168.0.0 is divisible by 256)
192.168.1.0/24 is valid (192.168.1.0 is divisible by 256)
192.168.0.128/24 is invalid (192.168.0.128 is not divisible by 256)

The alignment rule ensures that CIDR blocks don't overlap in unintended ways and can be aggregated cleanly.

Block Alignment Examples
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
# Understanding CIDR Block Alignment
# A block with prefix /n must start at an address divisible by 2^(32-n)
 
Example: /22 block (1,024 addresses)
Block size = 2^(32-22) = 2^10 = 1,024
 
Valid /22 blocks:
192.168.0.0/22   →  192.168.0.0 = 3232235520 (divisible by 1024 ✓)
192.168.4.0/22   →  192.168.4.0 = 3232236544 (divisible by 1024 ✓)
192.168.8.0/22   →  192.168.8.0 = 3232237568 (divisible by 1024 ✓)
 
Invalid /22 blocks:
192.168.1.0/22   →  192.168.1.0 = 3232235776 (not divisible by 1024 ✗)
192.168.2.0/22   →  192.168.2.0 = 3232236032 (not divisible by 1024 ✗)
192.168.3.0/22   →  192.168.3.0 = 3232236288 (not divisible by 1024 ✗)
 
Binary representation of valid /22 boundaries:
192.168.0.0   = 11000000.10101000.000000|00.00000000
192.168.4.0   = 11000000.10101000.000001|00.00000000
192.168.8.0   = 11000000.10101000.000010|00.00000000
                                        ^
                                The /22 boundary (last 10 bits = host portion)

The Binary Perspective

CIDR vs Classful: A Deep Comparison

Understanding the fundamental differences between classful and classless addressing clarifies why CIDR represented such a revolutionary change in Internet architecture.

Classful Addressing

•Network mask is implicit (determined by first bits)
•Only 3 network sizes: /8, /16, /24
•Wastes addresses in most allocations
•Every network = separate routing entry
•No hierarchical aggregation possible
•First octet determines network class
•Address space divided into 5 rigid classes
•Subnetting hidden from external routers

CIDR (Classless)

•Network mask is explicit (/n notation)
•Any network size from /1 to /32
•Allocations match actual needs precisely
•Multiple networks aggregated into one entry
•Hierarchical routing through aggregation
•First octet has no special meaning
•Address space treated as a continuum
•All prefix information propagated globally

The Routing Information Problem

Consider how routing changes between the two systems:

Classful Scenario: An ISP needs to advertise connectivity to 16 customers, each with a Class C network.

Routing table entries: 16 separate routes
Each route: 192.168.0.0, 192.168.1.0, ..., 192.168.15.0 (all /24)
No reduction possible

CIDR Scenario: The same 16 networks, if contiguous, become:

Routing table entries: 1 aggregated route
Single route: 192.168.0.0/20 (covers all 16 /24 networks)
93.75% reduction in routing table size

This aggregation capability is what saved the Internet from routing table collapse.

The CIDR Success Story

How CIDR Address Allocation Works

CIDR transformed not just the technical format of addresses, but the entire ecosystem of address allocation. Understanding this hierarchy is essential for network engineers.

The Hierarchical Allocation Model

CIDR enabled a clean hierarchical distribution of address space:

IANA (Internet Assigned Numbers Authority)
- Controls the entire IPv4 address space
- Allocates large blocks (/8 or larger) to Regional Internet Registries
RIRs (Regional Internet Registries)
- Five organizations covering global regions (ARIN, RIPE, APNIC, LACNIC, AFRINIC)
- Receive large blocks from IANA, allocate to ISPs and large organizations
- Typical allocations: /12 to /19
ISPs (Internet Service Providers)
- Receive blocks from RIRs
- Sub-allocate to customers
- Aggregate customer routes into single advertisements
- Typical allocations to customers: /24 to /29
End Organizations
- Receive address blocks from ISPs
- May further subnet internally
- Do not advertise to the global Internet directly

Converting Mermaid diagram...

Provider-Aggregatable (PA) vs Provider-Independent (PI) Addresses

CIDR creates two categories of address allocations:

Provider-Aggregatable (PA):

Addresses allocated from an ISP's block
Can be aggregated by the ISP (reducing global routing table)
If customer changes ISPs, addresses must be returned
Most common for typical organizations

Provider-Independent (PI):

Addresses allocated directly from an RIR
Cannot be aggregated (appear as separate routes)
Customer keeps addresses when changing ISPs
Reserved for organizations with strong justification
Contributes to routing table growth

The Aggregation Dependency

CIDR in Modern Networks

Cloud Computing and CIDR

Every major cloud provider uses CIDR extensively:

VPC/VNet design: Cloud virtual networks are defined using CIDR blocks
Subnet creation: VPCs are subdivided using smaller CIDR blocks
Security groups: Firewall rules reference CIDR blocks for source/destination
Route tables: All entries use CIDR notation

CIDR in Cloud Architecture (AWS Example)

Terraform

# AWS VPC Design Using CIDR Blocks
# Demonstrates hierarchical CIDR subdivision
 
resource "aws_vpc" "main" {
  cidr_block = "10.0.0.0/16"  # 65,536 addresses for entire VPC
  
  tags = {
    Name = "Production-VPC"
  }
}
 
# Public subnets: 10.0.0.0/20 - 10.0.15.255 (4,096 addresses each)
resource "aws_subnet" "public_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.0.0/20"   # 10.0.0.0 - 10.0.15.255
  availability_zone = "us-east-1a"
}
 
resource "aws_subnet" "public_az2" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.16.0/20"  # 10.0.16.0 - 10.0.31.255
  availability_zone = "us-east-1b"
}
 
# Private subnets: larger blocks for application servers
resource "aws_subnet" "private_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.128.0/18" # 10.0.128.0 - 10.0.191.255 (16K addresses)
  availability_zone = "us-east-1a"
}
 
resource "aws_subnet" "private_az2" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.192.0/18" # 10.0.192.0 - 10.0.255.255 (16K addresses)
  availability_zone = "us-east-1b"
}
 
# Database subnets: smaller, isolated blocks
resource "aws_subnet" "database_az1" {
  vpc_id     = aws_vpc.main.id
  cidr_block = "10.0.32.0/24"  # 10.0.32.0 - 10.0.32.255 (256 addresses)
  availability_zone = "us-east-1a"
}
 
# Security group using CIDR for access control
resource "aws_security_group_rule" "allow_internal" {
  type        = "ingress"
  from_port   = 0
  to_port     = 65535
  protocol    = "tcp"
  cidr_blocks = ["10.0.0.0/16"]  # Allow all traffic within VPC
  security_group_id = aws_security_group.internal.id
}
 
resource "aws_security_group_rule" "allow_partners" {
  type        = "ingress"
  from_port   = 443
  to_port     = 443
  protocol    = "tcp"
  cidr_blocks = [
    "203.0.113.0/24",    # Partner A network
    "198.51.100.0/25",   # Partner B network
  ]
  security_group_id = aws_security_group.api.id
}

Container Networking and CIDR

Container orchestration platforms like Kubernetes use CIDR extensively:

Pod CIDR: Each node receives a CIDR block for pod IP addresses
Service CIDR: Virtual IPs for services come from a dedicated block
Network Policies: Access control rules use CIDR notation

IPv6 and CIDR

CIDR concepts carry directly to IPv6, which was designed with CIDR from the start:

IPv6 prefix notation: 2001:db8::/32 (standard allocation size)
No classes ever existed in IPv6
Typically, enterprises receive a /48 (280 trillion subnets possible)
End sites often receive a /64 (standard subnet size)

CIDR Notation is Universal

Summary: The Classless Revolution

Key Takeaways

•CIDR eliminated class boundaries — The network/host division can occur at any bit position, not just at /8, /16, or /24.
•Prefix length is explicit — The /n notation clearly specifies the network portion, removing ambiguity.
•Block sizes follow powers of 2 — A /n block contains 2^(32-n) addresses, enabling predictable calculations.
•Alignment constraints apply — CIDR blocks must start at addresses divisible by their size.
•Hierarchical allocation enables aggregation — Contiguous allocations can be summarized into single routing entries.
•CIDR saved the Internet — Without it, address exhaustion and routing collapse would have occurred in the mid-1990s.

What's Next:

Page Complete

1 / 5