Week 7 · Lesson 6

Subnetting Across Many Networks (Fixed-Length)

A whole address block is assigned to your site. Now you must divide it into the many networks the design needs. This lesson does fixed-length subnetting end to end: split one block into equal pieces, list every subnet, and preview VLSM.

CCNA 200-301 Domain 1.6 & 1.7 IP Addressing Verified 2026-10-11
Part 1

Fixed-length subnetting: the rules

Fixed-length subnetting (FLSM) divides a block into subnets that are all the same size. You choose one borrowed-bit count and apply it everywhere. It is simple to plan and easy to document, which is why it appears in most exam design questions.

Definition — fixed-length rules. Given a parent block and a number of bits to borrow s:
  • Number of equal subnets = 2s.
  • Each subnet’s prefix = parent prefix + s (each subnet keeps the same size).
  • Each subnet’s address count = 2(32 − new prefix).

The whole calculation rests on two numbers: how many subnets you need, and how big each must be. Those pull in opposite directions — borrow more bits and you get more subnets, but smaller ones.

Fixed-length splitter · pick a parent and a target prefix

Part 2

Splitting a big block: 10.0.0.0/8

Suppose an ISP assigns your organisation 10.0.0.0/8 and you want one /16 per region. How many /16s can you make?

Parent: 10.0.0.0/8
Target: /16 per region
Borrowed s = 16 − 8 = 8 bits
Number of subnets = 2^8 = 256
Each /16 hosts = 2^16 − 2 = 65534

Subnet boundaries (octet 2 changes, increment 1):
  10.0.0.0/16     10.1.0.0/16     10.2.0.0/16    ...   10.255.0.0/16

Eight borrowed bits use the entire second octet: 10.n.0.0/16 for n = 0 to 255 is exactly 256 subnets, tiling the /8 with no gaps and no overlap. Pick any two: 10.10.0.0/16 runs from 10.10.0.0 to 10.10.255.255; 10.11.0.0/16 runs from 10.11.0.0 to 10.11.255.255. They are adjacent, never overlapping.

If instead you needed a /24 per department, borrow 24 − 8 = 16 bits to get 216 = 65,536 subnets of 254 hosts each — a good illustration of the trade-off. More, smaller subnets; or fewer, larger ones.

Exam tip. The relationship is always new prefix = parent prefix + borrowed bits, and subnets = 2borrowed. Do not mix the two: /8 to /16 borrows 8 bits and yields 28 = 256 subnets, not 16. The new prefix names the size, the borrowed count names the quantity.
Part 3

Full design: four /26s from a /24

A four-storey building is assigned 192.168.10.0/24. Each floor is its own network and needs up to 50 hosts. Follow the design.

Requirement: 4 subnets, each ≥ 50 hosts
Hosts: 50 -> need 2^h − 2 ≥ 50 -> h = 6 -> /26 (62 usable)
Borrowed s = 26 − 24 = 2  ->  2^2 = 4 subnets   (matches!)
Increment in octet 4 = 256 − 192 = 64
Boundaries: 0, 64, 128, 192
FloorSubnetNetworkFirstLastBroadcast
Floor 1/26192.168.10.0.1.62.63
Floor 2/26192.168.10.64.65.126.127
Floor 3/26192.168.10.128.129.190.191
Floor 4/26192.168.10.192.193.254.255

Verify Floor 3: network .128, broadcast .191, usable .129–.190 = 190 − 129 + 1 = 62 addresses. Each floor receives 62 usable, comfortably covering 50. The four blocks tile the /24 with no gaps, so no address is wasted beyond the two reserved per subnet.

Design check. Always confirm your two requirements pull together: the host requirement set the size (/26 = 62), and the floor count set the quantity (4 = 22). Here they agree exactly. If a design gave you 5 floors needing /26, a /24 could not hold 5 (only 4 fit), and you would need a second block or a smaller subnet — exactly the situation VLSM and summarisation solve.
Carving the /24 · pick the number of equal subnets

Part 4

The listing technique

To list every subnet quickly and without error, do not convert each address to binary. Instead:

  1. Compute the increment in the split octet.
  2. Start at the parent network’s value in that octet and add the increment repeatedly until you pass the octet’s maximum.
  3. For each value, the subnet is network at that value, broadcast at the next value minus one.

For the /26 design the increment is 64, so the octet-4 values are 0, 64, 128, 192 — four subnets, done. For a /27 the increment is 32 and the values are 0, 32, 64, 96, 128, 160, 192, 224 — eight subnets. The technique scales to any prefix and is far faster than binary on the clock.

Walk the block · step through the subnets of 192.168.10.0/24
Subnet boundaries in the last octet

Part 5

VLSM preview

Fixed-length subnetting is easy but wasteful when networks differ in size. Variable-Length Subnet Masking (VLSM) lets different subnets use different prefixes, sizing each to its need and packing them back-to-back. The detail is an advanced topic, but the idea is shown below.

Take 192.168.20.0/24 and four sites of 100, 50, 20, 10 hosts plus one router link of 2 hosts. Assign largest first:

PurposeNeedPrefixSubnetUsable rangeBroadcast
Sales100/25192.168.20.0/25.1 – .126.127
IT50/26192.168.20.128/26.129 – .190.191
Warehouse20/27192.168.20.192/27.193 – .222.223
WLAN10/28192.168.20.224/28.225 – .238.239
Router link2/30192.168.20.240/30.241 – .242.243

Check the packing: 128 + 64 + 32 + 16 + 4 = 244, so the first free address is .244 and .244–.255 (12 addresses) remain for growth. In a fixed-length scheme you would have had to use /25 everywhere, needing 5 × 128 = 640 addresses and overflowing the /24 entirely. VLSM fits all five networks into one /24 because it sizes each to need.

Calculation trick. The VLSM rule of thumb: place the largest subnets first, then the next largest, and so on down to the point-to-point /30s. Because each larger block is a multiple of every smaller block, they pack without gaps — the key property that makes variable sizing work.
Does the design fit? · fixed-length against variable-length
Fixed-length: every network takes a /25 (128 addresses)

Variable-length: each network sized to its own need

Part 6

A planning table

A design is not finished until it is written down. A useful plan names each subnet with its range and the devices it holds. Using the four-floor /26 design:

NameSubnetRangeDevices
Floor-1 LAN192.168.10.0/26.1–.62Access switches, PCs, printers
Floor-2 LAN192.168.10.64/26.65–.126Access switches, PCs, printers
Floor-3 LAN192.168.10.128/26.129–.190Access switches, PCs, printers
Floor-4 LAN192.168.10.192/26.193–.254Access switches, PCs, printers

The first usable address in each subnet is conventionally reserved for the default gateway (the router interface for that VLAN), which is why gateway addresses usually end in .1 (or .65, .129, .193 here). Documenting this now saves hours later.

Part 7

Key terms

Fixed-length subnetting (FLSM)

Dividing a block into equal-sized subnets.

Borrowed bits (s)

Bits taken from the host part to create more subnets.

Subnets = 2s

The number of equal subnets produced by borrowing s bits.

New prefix

Parent prefix + borrowed bits.

VLSM

Variable-Length Subnet Masking; subnets of different sizes packed to fit.

Block increment

256 − mask value in the split octet; the listing step.

Part 8 · Your turn

Review

Check what you have learned with these review questions.

Review · 6 questions

Sources

Sources & references

  1. Fuller, V. & Li, T. RFC 4632: Classless Inter-Domain Routing (CIDR) (2006) (block subdivision and VLSM background). rfc-editor.org/rfc/rfc4632.
  2. Pummill, R. & Manning, B. RFC 1878: Variable Length Subnet Table For IPv4 (1995) (the canonical VLSM/subnet table). rfc-editor.org/rfc/rfc1878.
  3. Rekhter, Y. et al. RFC 3021: Using 31-Bit Prefixes on IPv4 Point-to-Point Links (2000) (the /30 and /31 link masks). rfc-editor.org/rfc/rfc3021.
  4. Cisco. CCNA (200-301) Exam Topics, IPv4 Addressing, Subnetting, and VLSM. cisco.com.

Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.