Networking Fundamentals in 2026: IP, Subnet, DNS & IPv6
The moments when your code runs fine but "can't be reached" almost always trace back to a missing networking concept. Whether you're a developer or a sysadmin, knowing the basics—IP addresses, subnets, DNS, and ports—turns hours of debugging into minutes. This article is built as an up-to-date, practical reference for 2026. Less academic theory, more of what you'll actually use day to day.
What Is an IP Address? IPv4 and IPv6
An IP address is a device's identity on a network. Two versions are widely used side by side:
- IPv4: 32 bits, written as four octets:
192.168.1.10. It yields about 4.3 billion addresses, a pool that's long been exhausted. - IPv6: 128 bits, written as eight hexadecimal blocks:
2001:0db8:85a3::8a2e:0370:7334. In practice it offers a virtually unlimited supply (~3.4 × 10³⁸).
IPv4 scarcity has long been managed with NAT (Network Address Translation), hiding dozens of devices behind a single public IP. By 2026, IPv6 adoption has grown markedly on mobile and home internet worldwide; many large carriers now deploy IPv6 by default. IPv6 reduces the need for NAT, restores end-to-end connectivity, and is more efficient to route thanks to a simpler header. Still, IPv4 won't disappear any time soon—dual-stack setups that run both have become the norm.
Two shortcuts make IPv6 easier to write: a run of consecutive zero blocks can be collapsed once with ::, and leading zeros in each block can be dropped. So 2001:0db8:0000:0000:0000:0000:0000:0001 becomes 2001:db8::1.
Private vs Public IP Addresses
Not every IP is "visible on the internet." Certain ranges are reserved for private use and are only meaningful inside a local network:
| Range | CIDR | Typical use |
|---|---|---|
| 10.0.0.0 – 10.255.255.255 | 10.0.0.0/8 | Large enterprise networks |
| 172.16.0.0 – 172.31.255.255 | 172.16.0.0/12 | Mid-size networks, Docker |
| 192.168.0.0 – 192.168.255.255 | 192.168.0.0/16 | Home and office networks |
These addresses can't reach the internet directly; your router translates them to a single public IP via NAT. To quickly verify which class or type an address belongs to, use the IP Class Checker.
Subnet Masks and CIDR
An IP alone isn't enough; the subnet mask decides where the network part ends and the host (device) part begins. The mask is also 32 bits: from the left, 1-bits mark the network and 0-bits mark the host.
IP : 192.168.1.10
Mask : 255.255.255.0
11111111.11111111.11111111.00000000
└──────── network (24 bits) ─┘└ host ┘
CIDR notation expresses the mask as a single number like /24 instead of 255.255.255.0: it tells you how many bits go to the network. The number of usable hosts follows this formula:
Usable hosts = 2^(32 − prefix) − 2
The two subtracted addresses are the network address (all host bits 0) and the broadcast address (all host bits 1). Here are the values you'll meet most:
| CIDR | Subnet mask | Usable hosts |
|---|---|---|
| /30 | 255.255.255.252 | 2 |
| /29 | 255.255.255.248 | 6 |
| /28 | 255.255.255.240 | 14 |
| /26 | 255.255.255.192 | 62 |
| /24 | 255.255.255.0 | 254 |
| /16 | 255.255.0.0 | 65,534 |
Instead of working out a network's address, broadcast, and host range by hand, just feed an IP and prefix to the Subnet Calculator.
How DNS Works
Humans type magmanex.com; computers speak IP. The translation in between is handled by DNS (Domain Name System). The first time you reach a domain, resolution roughly goes like this:
- The browser and OS check their cache; if it's there, you're done.
- Otherwise the request goes to a resolver (typically your ISP or something like 1.1.1.1).
- The resolver starts at the root servers, asks the TLD servers (like
.com), then the domain's authoritative server. - The authoritative server returns the IP; the resolver caches it for the TTL duration.
This chain takes milliseconds, yet a single misconfigured record can make an entire site unreachable. To query a domain's records live, the DNS Lookup tool is handy.
DNS Record Types
| DNS record | Purpose |
|---|---|
| A | Maps a domain to an IPv4 address |
| AAAA | Maps a domain to an IPv6 address |
| CNAME | Points one domain to another (an alias) |
| MX | Specifies the domain's mail server and its priority |
| TXT | Free-form text; used for SPF, DKIM, and verification |
For example, the reason your email isn't being delivered is often a missing MX record or a wrong SPF (TXT) record; HTTPS certificate validations frequently ask for a TXT or CNAME record.
Ports: Same IP, Different Services
If an IP address is a building, a port is an apartment inside it. A server can run many services on a single IP at once, each separated by a different port number. A few worth knowing:
- 80 – HTTP
- 443 – HTTPS
- 22 – SSH
- 53 – DNS
- 5432 – PostgreSQL
"Connection refused" errors usually come from the wrong port, a service that isn't running, or a firewall rule. Always think of the ip:port pair together.
Summary
The core networking knowledge a developer should keep at hand in 2026:
- IPv4 is exhausted; IPv6 and dual-stack setups are now standard.
- Private IP ranges are local; they reach the internet via NAT.
- The CIDR prefix sets the network/host split: usable hosts = 2^(32 − prefix) − 2.
- DNS records (A, AAAA, CNAME, MX, TXT) serve different jobs; get one wrong and the service goes down.
- A port separates services living on the same IP.
Note: All networking tools on MagmaNex run entirely in your browser. No IP, domain, or network data you enter is ever sent to a server; no sign-up required, and it's free.

