Most companies are running out of IP addresses because they grow exponentially. They wrongly estimate the number of end devices that will be connected to the network. Someone says “about 500 addresses” and no one bothers to check, while in the same quarter a new office opens, a cloud project launches and a ton of IoT devices appear. Suddenly the network team has to rush to get a new block of IPs since IPv4 addresses are not the cheapest thing to purchase in bulk.
Luckily, there is no need to guess the required number of IPs. A simple 5-step method helps to calculate it.
Step 1: Count What You Use Today
Start with the inventory. Take a look at all the IPs your company is using right now:
• Servers and virtual machines
• Your employees' personal devices
• Printers, cameras and other equipment
• Public services (websites, mail servers, VPN gateways)
• Cloud and hosting platforms
• Customer addresses, for ISPs or hosting companies
IP addresses should be pulled from your IPAM tool, DHCP servers or network sweeps. Forget about the spreadsheet you saved last year, it's most likely out of date.
Step 2: Separate "Assigned" from "Actually Used"
Many companies have addresses that nobody is using. Old test servers, retired devices and forgotten subnets.
Record two numbers. How many addresses are assigned and how many are actually being used. The difference between these two numbers is space you may be able to reclaim before having to buy anything new.
Step 3: Estimate Your Growth
This is the section that people tend to overlook. I want you to consider three aspects and express them as a percentage:
• Growth: What has been the growth of your address base during the past 12 months? For instance, have you added 1000 new addresses per year or 200?
• Plans: Will you be launching new products/offices this year? Are there any plans for expansion or acquisition of new customers in your address book?
• Technology: Are you embracing more remote employees, cloud-based software, and IoT-enabled devices?
Step 4: Add a Safety Buffer
**Never plan on designing for exactly what you think you'll need. Always budget an additional 10-20% for things like unexpected projects or new customers who need a chunk of address space allocated to them.
You also have to consider subnetting waste where addresses are lost to network and broadcast addresses and when you divide a subnet into smaller subnets of different sizes, there will always be some addresses left over that can't be used. Another 10-15% is a good rule of thumb to use.
Step 5: Round Up to a Real Block Size
IP space comes in fixed sizes, so your final number must match one. Here are the common ones:
Block Total Addresses
/24 256
/23 512
/22 1,024
/21 2,048
/20 4,096
A Quick Example
Let's say your company uses 1,200 addresses today.
• Expected growth (20%): +240, so 1,440
• Safety buffer (15%): +216, so about 1,656
• Nearest block that fits: a /21 (2,048 addresses)
A /22 would have felt like enough, but it would have run out in months. This is exactly why rounding matters.
Don't Forget This: Regulators May Ask You to Prove It
If your plan is to acquire space via RIR transfer, then the registry will most likely require a justification for the need of the resources and that need should be for some period of time, often 12-24 months as a rough estimate (but it varies) based on the registry and something concrete like the calculation above will definitely make the process a lot easier compared to just throwing darts at the wall and hoping for the best.
Common Mistakes to Avoid
• Only considering devices, and not servers, gateways and infrastructure
• Ignoring NAT and cloud use that quietly eats up public addresses
• Planning for one year, but buying for exactly one year, with no buffer
• Never auditing old subnets to look for space
Top comments (0)