A port helps identify a transport endpoint on a host. Common assignments are useful clues, but applications can use different ports. A port number or an open result does not verify which software is listening or whether access is safe.
A practical port reference
| Port | Common use | Typical transport/context |
|---|---|---|
| 22 | SSH | TCP; administration and secure shell services |
| 25 | SMTP relay | TCP; server-to-server mail |
| 53 | DNS | UDP and TCP |
| 80 | HTTP | TCP; unencrypted HTTP or redirect |
| 123 | NTP | UDP; time synchronization |
| 443 | HTTPS | TCP for common HTTP/TLS; UDP for QUIC/HTTP/3 |
| 445 | SMB | TCP; file-sharing services |
| 465 / 587 | Mail submission | TCP; implicit TLS / submission, respectively |
| 853 | DNS over TLS | TCP for DoT |
| 993 | IMAP over TLS | TCP; mail retrieval |
| 3389 | Remote desktop | TCP and UDP depending on configuration |
Use the service's official requirements for real configuration. Some protocols use additional negotiated ports or different transport arrangements. A compact reference cannot replace a product's deployment documentation.
Distinguish a label from a verified service
A custom application could listen on TCP 22 without being SSH. A website could use HTTPS on 8443. Tools may display a conventional service name from a port table even when they have not identified the application.
TCP and UDP port numbers are separate. A TCP check on 53 does not prove that UDP DNS queries work. A timeout can reflect filtering, a missing listener, a stopped service, routing trouble, or probe behavior; it is not a detailed vulnerability finding.
Separate local use from public exposure
A service needed within a trusted LAN may not belong on the public Internet. Keep file sharing and administration behind appropriate private access boundaries. If you intentionally publish a service, scope the firewall rule, require authentication, keep software supported, and monitor its own logs.
Using an unusual port may reduce incidental noise but does not replace those controls. A broad rule can expose more than the one service you meant to test, and automatically created mappings should be reviewed.
Use port information in a diagnostic sequence
- Identify the intended application, transport, and port.
- Confirm the actual listener and local connectivity.
- Check host and router rules, plus any upstream NAT or ISP filter.
- Test externally with the correct protocol and the service's supported diagnostic.
This site's port checker attempts only selected TCP ports on your observed public IP. Its output is a useful reachability observation, not permission to scan another network or a certification that your server is secure.
Put it into practice
References & further reading
Protocol standards and official documentation supporting this guide. Device menus and provider policies vary; use the documentation for your equipment when applying a change.