/21 vs /22 — Subnet Comparison

A /21 subnet is larger than a /22. Every additional bit in the prefix halves the address space — the 1-bit difference between these two means /21 has 21 = 2 times as many addresses.

/21

2K IPs — building-scale subnet

Full reference →
Total IPs 2,048
Usable Hosts 2,046
Subnet Mask 255.255.248.0
Wildcard Mask 0.0.7.255

Typical Uses

  • Large application tier
  • Enterprise building network
  • Kubernetes node pool subnet
/22

1K IPs — medium site subnet

Full reference →
Total IPs 1,024
Usable Hosts 1,022
Subnet Mask 255.255.252.0
Wildcard Mask 0.0.3.255

Typical Uses

  • Medium office floor VLAN
  • Application tier with ~500 hosts
  • Cloud subnet for a single microservice cluster

Key Differences

more IPs in /21 than /22
2
/22 subnets fit inside one /21
1
bit of difference in prefix length

How 2 /22 Subnets Divide a /21

Example using 10.0.0.0/21 as the parent block.

# CIDR Network First Usable Last Usable Broadcast Hosts
1 10.0.0.0/22 10.0.0.0 10.0.0.1 10.0.3.254 10.0.3.255 1,022
2 10.0.4.0/22 10.0.4.0 10.0.4.1 10.0.7.254 10.0.7.255 1,022

FAQ

What is the difference between /21 and /22?

A /21 has 2,046 usable hosts and a /22 has 1,022. The subnet masks differ: /21 uses 255.255.248.0 while /22 uses 255.255.252.0. Every additional bit in the prefix halves the number of addresses — so the 1-bit gap means /21 is exactly 2× larger.

How many /22 subnets fit in a /21?

Exactly 2 /22 subnets fit perfectly inside one /21 with no wasted space. To split a /21 into /22s, just increment the last 1 bit of the network address for each new subnet.

Which should I choose?

/21 is typically used for: Enterprise building / large app tier. /22 is better for: Medium office/application segment. Choose the smallest prefix that comfortably fits your host count — over-allocating wastes address space, but under-allocating means painful renumbering later.

Related Comparisons