| Metric | Value | Copy |
|---|---|---|
| {{ row.key }} | {{ row.value }} |
| Offset | Address | Role | Copy |
|---|---|---|---|
| {{ row.row }} | {{ row.address }} | {{ row.role }} |
| Prefix | Mask | Wildcard | Total | Usable | Relative size | Host target fit |
|---|---|---|---|---|---|---|
| {{ row.prefix }} | {{ row.mask }} | {{ row.wildcard }} | {{ row.totalAddresses }} | {{ row.usableHosts }} | {{ row.relativeSize }} | {{ row.fitLabel }} |
| # | Demand | Planned | Capacity | CIDR | First | Last | Broadcast | Waste | Status |
|---|---|---|---|---|---|---|---|---|---|
| {{ row.order }} | {{ row.requestedDisplay }} | {{ row.plannedDisplay }} | {{ row.capacityDisplay }} | {{ row.cidr }} | {{ row.firstHost }} | {{ row.lastHost }} | {{ row.broadcast }} | {{ row.wasteDisplay }} | {{ row.status }} |
| Topic | Value | Copy |
|---|---|---|
| {{ row.key }} | {{ row.value }} |
IPv4 subnetting answers a practical question: given one address and one boundary, which addresses belong to the same network and which do not? That answer drives VLAN sizing, interface configuration, firewall scope, ACL rules, VPN handoffs, DHCP pools, and route summaries. A small mistake in prefix length can push traffic into the wrong network, hide an overlap, or make a host range look larger than it really is.
This calculator starts from the formats people usually receive in tickets and design notes. You can enter CIDR such as 192.168.14.99/26, an address plus dotted mask such as 192.168.14.99 255.255.255.192, or even a bare IPv4 address if you want to test it with the default /24. The result is expanded into the network address, subnet mask, wildcard mask, first and last usable hosts, broadcast address, total addresses, usable hosts, previous and next peer subnet, and a short address-type summary.
The tool also covers the parts of subnet planning that usually get scribbled in margins. You can test a host target against the current subnet, compare nearby prefixes in the Prefix Ladder, inspect a 0-based host offset in the Host Window, and build a reserve-adjusted VLSM split plan from a list of demand sizes. Each result view can be copied or downloaded as CSV, exported as DOCX, or captured as structured JSON for documentation work.
All subnet math happens in the browser. That is useful for quick checks on internal ranges, but there is one sharing detail to remember: changed inputs are reflected in the page URL so the share link can reopen the same scenario later. Treat copied links accordingly if the address plan is sensitive.
An IPv4 address has 32 bits. A prefix length such as /26 says how many leading bits belong to the network portion. The remaining bits belong to hosts inside that network. CIDR made this explicit bit boundary more important than the older classful A, B, and C labels, which is why the calculator reports both the working prefix and the legacy class family but treats the prefix as the source of truth.
From the chosen prefix, the calculator builds the subnet mask, inverts that mask to produce the wildcard mask, applies the mask to the entered address to find the network base, and fills all host bits with ones to find the broadcast value. The Address Math tab surfaces the same result as dotted decimal, hexadecimal, bit strings, integer span, and an ACL-style wildcard line so you can audit the boundary rather than trust a summary alone.
192.168.14.99/26 11000000 10101000 00001110 | 01100011 network bits: 26 | host bits: 6
Host counts follow the prefix, not habit. For ordinary subnets, usable hosts equal 2^(32 - prefix) - 2 because the all-zero host value is the network address and the all-one host value is the broadcast address. Two edge cases matter. A /31 is treated as a point-to-point link with two usable endpoints, following RFC 3021. A /32 is treated as a single host route. Those two prefixes are why the Host Window and summary line look different near the top of the prefix range.
| Stage | What the tool does | What you see |
|---|---|---|
| Demand sizing | Each requested host count is increased by the reserve percentage and rounded up. | Planned can be larger than Demand. |
| Prefix choice | The tool picks the smallest standard prefix whose usable-host capacity fits the planned load. | Capacity and CIDR show the fitted block size. |
| Alignment | Each child block starts on its natural binary boundary inside the parent range. | Largest-first ordering usually packs better than small-first ordering. |
| Failure check | If a child block would need a larger parent or would run past the parent end, the row is rejected. | Status becomes Needs larger parent or Insufficient address space. |
The address-type badge is a convenience layer, not a routing verdict. The tool checks the entered address against a built-in list of notable IPv4 ranges and labels matches such as private use, shared CGN space, loopback, link-local, documentation-only ranges, benchmarking, multicast, and reserved space. If nothing matches, the tool shows Public / unclassified address space. That means only that the address was not found in the built-in special-use list. It does not prove public reachability, ownership, delegation, or safe deployment.
| Range | Meaning in practice |
|---|---|
10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16 |
RFC 1918 private-use space for internal networks. It is not globally routed on the public Internet. |
100.64.0.0/10 |
Shared CGN address space. It is special-use carrier space, not a normal enterprise private block. |
127.0.0.0/8 and 169.254.0.0/16 |
Loopback and link-local ranges used for local stack testing or same-link fallback, not routed subnet design. |
192.0.2.0/24, 198.51.100.0/24, 203.0.113.0/24 |
Documentation-only TEST-NET ranges meant for examples, books, and lab-style illustrations. |
198.18.0.0/15, 224.0.0.0/4, 240.0.0.0/4 |
Benchmarking, multicast, and reserved space. These ranges should not be treated as ordinary host addressing. |
Start with the notation you were given, not the notation you think it should have been. If a request says 10.12.48.77 255.255.255.224, paste that exact pair first. The calculator will normalize it into CIDR and show whether the expected network, host range, and broadcast really match the handoff. That is the fastest way to catch a copied mask, an off-by-one host assumption, or a gateway assigned outside the usable range.
Use Prefix override only when you are deliberately testing a different boundary against the same IPv4 address. It is useful when someone sends a host address without a prefix, or when you want to compare how the same address behaves as /24, /25, and /26. It is not a repair button for a bad input. The first pass should usually be Auto so you can see what the original notation actually implies.
When you are sizing a live subnet, set Host target before you look at the Prefix Ladder. The calculator applies the reserve percentage to that target and then tells you whether the current block fits, how much spare room remains, and what the smallest fitting prefix would be. That is more useful than eyeballing raw host counts because a reserve bump often moves the design into the next larger block.
When you are slicing a parent network into child networks, the split planner deserves slower reading. Demand is the number you entered. Planned is the reserve-adjusted demand. Capacity is the usable-host count for the chosen child prefix. Waste shows spare usable host slots inside that child subnet, not the total raw addresses left in the parent. Largest-first ordering is usually best when your goal is efficient packing. Keep-entered-order is better when you are staging a rollout and sequence matters more than packing efficiency.
The Host Window is best used as a sampler, not a full IPAM list. The offset is 0-based across usable hosts, so probe index 0 is the first usable address in a normal subnet. The focus mode can show the start of the range, center around the chosen probe, or jump to the tail of the subnet. For large blocks, the window can hide earlier or later rows and tell you how many were skipped. That keeps the page readable while still letting you inspect a precise position.
The export options map to common operations work. Deploy Brief, Host Window, Prefix Ladder, Split Plan, and Address Math can each be copied as CSV, downloaded as CSV, or exported as DOCX. The JSON tab is the compact structured view if you need to hand the result to another person or script. Because the page state is mirrored into the URL, the share link is convenient for peer review but should be handled carefully on private address plans.
/24.The most important output is still the network boundary. If the CIDR, first host, last host, or broadcast do not match the intended subnet, everything else on the page is describing the wrong block.
| Tab | Question it answers |
|---|---|
| Deploy Brief | What subnet did this input actually describe, and does it fit the host target or split plan summary? |
| Host Window | Which exact addresses sit at the beginning, around a probe offset, or near the end of the usable range? |
| Prefix Ladder | How do nearby prefixes compare in total size, usable hosts, wildcard, and fit against the planned load? |
| Split Plan | Can this parent subnet hold the reserve-adjusted child demands, and where would each child block land? |
| Address Math | What are the binary, hexadecimal, integer-span, and ACL wildcard representations of the same subnet? |
| JSON | What does the full result look like in a structured format for logging, review, or automation? |
Class family is there for orientation only. It is derived from the first octet and can be useful when reading old documentation, but it does not control the actual subnet calculation. Address type is also contextual rather than authoritative. A private-use label tells you the range belongs to RFC 1918 space, and a public or unclassified label tells you only that the entered address was not matched to one of the built-in special-use blocks.
Previous subnet and Next subnet are good overlap checks. They help you see whether the current block sits where you expected in a larger summary range. The ACL wildcard line is useful when you are moving from subnet planning into access-list work and need the wildcard representation without redoing the inversion by hand.
In the split plan, a green Allocated badge means only that the child block fit inside the current parent under the selected reserve and ordering rules. It does not mean the block is free in your wider IP plan. A red failure status means something concrete: either the child demand requires a larger parent prefix, or earlier allocations have already consumed enough aligned space that the remaining part of the parent can no longer hold the next block.
192.168.14.99/26. The calculator resolves that to network 192.168.14.64/26, first usable host 192.168.14.65, last usable host 192.168.14.126, broadcast 192.168.14.127, and 62 usable hosts. If a ticket says the gateway is 192.168.14.130, the Deploy Brief makes the mistake obvious because that address sits outside the computed host range.203.0.113.10/31. The summary line switches to a two-point link view and the Host Window shows two usable endpoints: 203.0.113.10 and 203.0.113.11. There is no ordinary network-plus-broadcast host exclusion here, which is exactly why /31 is useful on point-to-point links and confusing on a quick glance if you expect LAN-style rules.10.0.8.0/24 and set the split demand list to 60, 30, 12 with a 10% reserve. The reserve-adjusted demands become 66, 33, and 14 usable hosts, so the planner chooses 10.0.8.0/25, 10.0.8.128/26, and 10.0.8.192/28. All three fit, and the split summary reports 48 raw addresses still unassigned in the parent block.Because the calculator treats /31 as a point-to-point link. RFC 3021 allows the two values in that subnet to be used as endpoints rather than reserving one as a network address and the other as a broadcast address.
The calculator defaults it to /24 unless you set a prefix override. That makes testing easy, but it also means a bare address should be treated as a draft assumption until you confirm the real prefix.
No. The dotted mask has to represent contiguous network bits. A non-contiguous mask is rejected instead of being guessed or normalized.
It means the entered address did not match one of the built-in special-use ranges checked by the tool. It is not proof that the block is routed on the public Internet or assigned to your organization.
The subnet calculations, host sampling, prefix comparison, and split planning run in the browser. The main privacy caveat is that changed inputs are mirrored into the URL so the share link can reopen the same scenario.
Large blocks have fewer valid alignment positions. Allocating them first usually reduces fragmentation and makes it easier to fit smaller child networks into the remaining gaps.