Loading learning content...
You've been given an IP address block: 172.16.0.0/16. Your organization has 12 departments ranging from 20 to 500 employees, a data center requiring 100 server addresses, 15 point-to-point WAN links, and anticipated 50% growth over five years. How do you divide this address space?
This is subnet creation—the process of transforming business requirements into a logical, efficient, and scalable IP addressing architecture. It's part mathematics, part engineering judgment, and part organizational understanding.
By the end of this page, you will understand: (1) The systematic process for subnet design, (2) How to determine optimal subnet sizes, (3) Techniques for allocating subnets to organizational units, (4) Best practices for addressing scheme documentation, and (5) Common pitfalls and how to avoid them.
Subnet creation follows a systematic process that balances technical constraints with organizational requirements. The steps build upon each other—rushing through early steps creates problems in later phases.
Renumbering an IP addressing scheme after deployment is extremely disruptive and expensive. DNS updates, firewall rules, application configurations, documentation, and user training all require changes. Invest time in proper design upfront—it's far cheaper than remediation.
Before touching any IP addresses, you need a complete picture of what the network must support. This involves stakeholder interviews, documentation review, and future planning.
Example Requirements Document:
Organization: Acme Corporation
Total Employees: 2,500 (current), 4,000 (projected 5-year)
Departments:
├── Engineering (450 employees, expected to double)
├── Sales (300 employees)
├── Marketing (150 employees)
├── Finance (100 employees, strict isolation required)
├── HR (75 employees, strict isolation required)
├── Operations (200 employees)
├── Executive (50 employees)
├── IT (125 employees)
└── Guest/Contractor (variable, up to 200 concurrent)
Infrastructure:
├── Data Center (150 servers, expandable to 300)
├── DMZ (25 public-facing servers)
├── Management Network (50 network devices)
├── VoIP (2,500 phones, same growth as employees)
└── IoT/Building Systems (500 devices)
WAN Links: 8 point-to-point connections
Allocated Address Space: 10.0.0.0/8 (private)
Once requirements are documented, translate host counts into subnet sizes. The fundamental formula:
Usable Hosts = 2^(32 - prefix) - 2
The -2 accounts for the network address (all-zeros host portion) and broadcast address (all-ones host portion), which cannot be assigned to devices.
Sizing for Requirements:
Always round UP to the next power of 2, then add growth buffer:
| Host Requirement | Choose Prefix | Actual Capacity | Efficiency* |
|---|---|---|---|
| 2 | /30 | 2 | 100% |
| 3-6 | /29 | 6 | 50-100% |
| 7-14 | /28 | 14 | 50-100% |
| 15-30 | /27 | 30 | 50-100% |
| 31-62 | /26 | 62 | 50-100% |
| 63-126 | /25 | 126 | 50-100% |
| 127-254 | /24 | 254 | 50-100% |
| 255-510 | /23 | 510 | 50-100% |
| 511-1,022 | /22 | 1,022 | 50-100% |
When current utilization exceeds 50% of subnet capacity, it's time to plan for expansion. This buffer handles unexpected growth from new projects, acquisitions, or IoT proliferation. Never design subnets to be more than 50% utilized at deployment.
Applying to Example Requirements:
| Subnet Purpose | Current Hosts | With Growth | Prefix | Capacity |
|---|---|---|---|---|
| Engineering | 450 | 900 | /22 | 1,022 |
| Sales | 300 | 450 | /23 | 510 |
| Marketing | 150 | 225 | /24 | 254 |
| Finance | 100 | 150 | /24 | 254 |
| HR | 75 | 120 | /25 | 126 |
| Operations | 200 | 300 | /23 | 510 |
| Executive | 50 | 75 | /25 | 126 |
| IT | 125 | 190 | /24 | 254 |
| Guest | 200 | 300 | /23 | 510 |
| Data Center | 150 | 300 | /23 | 510 |
| DMZ | 25 | 50 | /26 | 62 |
| Management | 50 | 100 | /25 | 126 |
| VoIP | 2,500 | 4,000 | /20 | 4,094 |
| IoT | 500 | 1,000 | /22 | 1,022 |
| WAN Links | 8 | 16 | 16×/30 | 2 each |
Total Address Consumption Analysis:
Summarizing the above allocations:
Approximate total: ~10,500 addresses Available in 10.0.0.0/8: 16,777,214 addresses
✓ Requirements fit comfortably within allocated space.
With subnet sizes determined, the next step is assigning specific address ranges. This isn't arbitrary—good allocation strategies simplify routing, aid troubleshooting, and enable future summarization.
Hierarchical Allocation Example:
Given address space: 10.0.0.0/8
First, divide into major categories using the second octet:
10.0.0.0/16 - Reserved for future
10.1.0.0/16 - WAN/Infrastructure
10.2.0.0/16 - Reserved for future
10.10.0.0/16 - User Networks (Departments)
10.11.0.0/16 - User Networks (Expansion)
10.20.0.0/16 - Voice/VoIP
10.30.0.0/16 - Data Center/Servers
10.40.0.0/16 - Security Zones (DMZ, Guest)
10.50.0.0/16 - IoT/OT Networks
10.100-254 - Reserved for future
Detailed User Network Allocation (10.10.0.0/16):
10.10.0.0/22 - Engineering (1,022 hosts)
10.10.4.0/23 - Sales (510 hosts)
10.10.6.0/24 - Marketing (254 hosts)
10.10.7.0/24 - Finance (254 hosts, isolated)
10.10.8.0/25 - HR (126 hosts, isolated)
10.10.8.128/25 - Executive (126 hosts)
10.10.9.0/24 - IT (254 hosts)
10.10.10.0/23 - Operations (510 hosts)
10.10.12.0/22 - [Reserved for growth]
Note the Alignment:
This alignment isn't arbitrary—it's required. A /22 contains 4 consecutive /24s, so it must start at a multiple of 4.
Subnets must be aligned to their natural boundaries. A /22 (1,024 addresses) must start at an address divisible by 1,024. A /27 (32 addresses) must start at an address divisible by 32. Misalignment causes overlapping ranges and routing failures.
Let's work through a complete subnet creation exercise to solidify the process.
Scenario:
You're designing the IP scheme for a new branch office with:
192.168.100.0/23 (512 addresses)Step-by-Step Solution:
1. Calculate Required Sizes:
| Subnet | Current | +50% Growth | Min Size | Chosen Prefix | Addresses Used |
|---|---|---|---|---|---|
| Engineering | 45 | 68 | /25 (126) | /25 | 128 |
| Sales | 28 | 42 | /26 (62) | /26 | 64 |
| Admin | 12 | 18 | /27 (30) | /27 | 32 |
| Servers | 8 | 12 | /28 (14) | /28 | 16 |
| Printers/IoT | 15 | 23 | /27 (30) | /27 | 32 |
| Guest | 50 | 75 | /25 (126) | /25 | 128 |
| WAN Link 1 | 2 | 2 | /30 (2) | /30 | 4 |
| WAN Link 2 | 2 | 2 | /30 (2) | /30 | 4 |
Total Required: 128+64+32+16+32+128+4+4 = 408 addresses Available: 512 addresses in /23 Remaining: 104 addresses (for future needs) ✓
2. Allocate Addresses (Largest First):
Base range: 192.168.100.0 - 192.168.101.255
/25 blocks (128 addresses each):
192.168.100.0/25 → Engineering (100.0 - 100.127)
192.168.100.128/25 → Guest (100.128 - 100.255)
/26 blocks (64 addresses each):
192.168.101.0/26 → Sales (101.0 - 101.63)
/27 blocks (32 addresses each):
192.168.101.64/27 → Admin (101.64 - 101.95)
192.168.101.96/27 → Printers/IoT (101.96 - 101.127)
/28 blocks (16 addresses each):
192.168.101.128/28 → Servers (101.128 - 101.143)
192.168.101.144/28 → [Reserved]
/30 blocks (4 addresses each):
192.168.101.160/30 → WAN Link 1 (101.160 - 101.163)
192.168.101.164/30 → WAN Link 2 (101.164 - 101.167)
Remaining: 192.168.101.168 - 192.168.101.255 (88 addresses reserved)
| Subnet Name | Network Address | Mask | Usable Range | Broadcast | Gateway* |
|---|---|---|---|---|---|
| Engineering | 192.168.100.0 | /25 | .1 - .126 | .127 | .1 |
| Guest | 192.168.100.128 | /25 | .129 - .254 | .255 | .129 |
| Sales | 192.168.101.0 | /26 | .1 - .62 | .63 | .1 |
| Admin | 192.168.101.64 | /27 | .65 - .94 | .95 | .65 |
| Printers/IoT | 192.168.101.96 | /27 | .97 - .126 | .127 | .97 |
| Servers | 192.168.101.128 | /28 | .129 - .142 | .143 | .129 |
| WAN Link 1 | 192.168.101.160 | /30 | .161 - .162 | .163 | N/A |
| WAN Link 2 | 192.168.101.164 | /30 | .164 - .165 | .167 | N/A |
The gateway address (usually the router's interface in that subnet) is conventionally the first (.1) or last (.254 for /24) usable address. Consistency across the organization aids troubleshooting. Document whichever convention you choose.
A subnetting scheme is only as good as its documentation. Poorly documented IP schemes lead to conflicts, confusion, and costly mistakes during troubleshooting or expansion.
12345678910111213141516171819202122232425262728293031323334353637383940414243444546
# Branch Office IP Addressing Scheme# Document Version: 1.0# Last Updated: 2024-01-15# Author: Network Engineering Team site: name: "Branch Office - Seattle" allocated_block: "192.168.100.0/23" total_addresses: 512 allocated_addresses: 408 reserved_addresses: 104 subnets: - name: "Engineering" network: "192.168.100.0/25" vlan_id: 100 gateway: "192.168.100.1" usable_range: "192.168.100.2 - 192.168.100.126" dhcp_scope: "192.168.100.10 - 192.168.100.126" broadcast: "192.168.100.127" purpose: "Engineering department workstations" reserved_ips: - ip: "192.168.100.1" purpose: "Default gateway" - ip: "192.168.100.2" purpose: "Dept printer" - name: "Servers" network: "192.168.101.128/28" vlan_id: 200 gateway: "192.168.101.129" usable_range: "192.168.101.130 - 192.168.101.142" dhcp_scope: null # Static only purpose: "Branch office servers" reserved_ips: - ip: "192.168.101.130" hostname: "sea-dc01" purpose: "Domain Controller" - ip: "192.168.101.131" hostname: "sea-fs01" purpose: "File Server" unallocated: - range: "192.168.101.168 - 192.168.101.255" addresses: 88 intended_purpose: "Future department expansion"Even experienced network engineers make subnetting mistakes. Awareness of common pitfalls helps avoid them.
Overlapping subnets cause intermittent, hard-to-diagnose connectivity issues. For example, 192.168.1.0/24 and 192.168.1.0/25 overlap—the /25 is a subset of the /24. If both are configured on different routers, packets may route incorrectly depending on which route is preferred. Always verify no overlap exists before deploying.
Overlap Detection Technique:
Two subnets overlap if and only if one's network address falls within the other's range (or vice versa).
Subnet A: 192.168.100.0/25 (range: .0 - .127)
Subnet B: 192.168.100.64/26 (range: .64 - .127)
Check: Does 192.168.100.64 fall within 192.168.100.0 - 192.168.100.127?
Yes → OVERLAP! These subnets cannot both exist.
Subnet C: 192.168.100.128/25 (range: .128 - .255)
Check: Does 192.168.100.128 fall within Subnet A (.0 - .127)?
No → No overlap. These can coexist.
Automate this check with IP address management (IPAM) tools for large environments.
Subnet creation is the translation of organizational requirements into a logical, efficient IP addressing architecture. It requires both technical precision and organizational understanding.
What's Next:
With the subnet creation process understood, we need to master the mathematical calculations that underpin it. The next page covers subnet calculations—the formulas and techniques for determining network addresses, broadcast addresses, valid host ranges, and performing rapid mental math for subnetting questions.
You now understand the complete subnet design process from requirements to documentation. You can translate organizational needs into subnet specifications and allocate address space efficiently. Next, we'll master the calculations that make this work precisely.