Protocol stack and layering
RDP is a layered, multi-channel protocol. At its base it runs over TCP (and, since RDP 8, UDP for performance-sensitive graphics traffic). Above transport sit the security and session layers: connection negotiation, TLS encryption, and authentication happen before any desktop data flows. The session layer then multiplexes several logical streams — graphics output, user input, and a set of virtual channels for everything else.
Microsoft built RDP on concepts from the ITU-T T.120 family of conferencing standards, extending them heavily over the years. The practical upshot of the layering: each concern — transport reliability, encryption, graphics encoding, device redirection — can evolve independently, which is how a protocol from the 1990s still feels modern.
Transport: TCP, UDP, and port 3389
RDP's well-known port is 3389 — TCP traditionally, with UDP 3389 added for RDP 8 and later to carry graphics data with lower latency on lossy networks. The client initiates the connection to the server's port; after capability negotiation, the session upgrades through its security handshake.
Port 3389 is so synonymous with RDP that it's a fixture in firewall rules, security scanners, and attack botnets alike. Administrators commonly change the listening port, restrict it by source IP, or front it with a VPN or RD Gateway — not because the protocol is broken, but because a well-known port attracts automated probing. Our RDP port guide and port 3389 guide cover the practical side.
Encryption and authentication
Early RDP used weaker native encryption; modern deployments should use TLS with Network Level Authentication (NLA). NLA — built on CredSSP — authenticates the user before the server allocates session resources, which both protects credentials and blunts denial-of-service and pre-authentication attacks. Certificates validate the server's identity; clients should be configured to warn on or reject untrusted certificates rather than clicking through blindly.
Authentication itself typically flows through Windows credentials (username/password, smart cards, or integrated Windows authentication), and can be layered with multi-factor requirements via RD Gateway or VPN front-ends. The protocol supports the mechanisms — the deployment decides how strictly to use them.
Virtual channels
Virtual channels are RDP's extension system: named, bidirectional pipes multiplexed inside the session, each carrying one class of data. Standard channels handle clipboard, drive/printer/port redirection, audio, and smart cards; applications and administrators can add custom channels. This is what makes an RDP session feel integrated with the local machine rather than isolated from it.
Channels are also a security boundary worth managing. Each redirected resource — a local drive, a printer, the clipboard — is a bridge between the remote session and the local device. Hardened deployments disable redirection they don't need, especially on servers reachable from less-trusted networks.
Major versions at a glance
| Version | Era / shipped with | Notable additions |
|---|---|---|
| RDP 5.x | Windows 2000 / XP era | Foundation of modern RDP sessions |
| RDP 6.x | Vista / Server 2008 era | Improved security, multi-monitor groundwork |
| RDP 7 | Windows 7 / Server 2008 R2 | Multimedia and graphics redirection improvements |
| RDP 8 / 8.1 | Windows 8 / Server 2012 era | UDP transport, RemoteFX graphics, adaptive graphics |
| RDP 10 | Windows 10 / Server 2016+ | Current generation; H.264/AVC graphics, modern auth |
Version negotiation happens automatically: client and server agree on the highest version both support. Keeping both ends updated is the simplest way to get the protocol's best performance and security.
Hardening checklist
Whether you run your own server or rent one, these are the non-negotiables:
- Enforce NLA and TLS. Don't allow legacy encryption fallbacks.
- Strong, unique credentials plus an account lockout policy against brute force.
- Patch promptly. Protocol-level fixes only protect patched machines.
- Limit exposure. IP allowlists, VPN, or RD Gateway instead of internet-wide 3389.
- Trim virtual channels. Disable drive, printer, and clipboard redirection you don't use.
- Monitor logons. Review security logs for failed logon bursts — the signature of automated probing.
RDP's reputation for insecurity comes almost entirely from deployments that skipped this list. The protocol itself, configured to modern standards, is a mature and defensible remote-access choice — which is why it powers everything from enterprise infrastructure to rented RDP servers.
Ready for your own RDP server?
Full admin access, dedicated IP, 24/7 support and a 3-day money-back guarantee on every plan. Pay by credit/debit card, bank transfer, Bitcoin or altcoins.
3-day money-back guarantee · Instant setup · 30+ server locations · 99% uptime SLA
RDP Protocol: FAQ
The RDP protocol is Microsoft's technical standard that lets one computer display and control another computer's desktop over a network. It defines how screen images, keyboard/mouse input, and extras like clipboard and audio travel between the two machines, all encrypted.
RDP is proprietary to Microsoft, though its specifications are publicly documented. It builds on concepts from the ITU-T T.120 conferencing standards and has been extended through many versions since the late 1990s.
Modern RDP uses TLS encryption combined with Network Level Authentication (NLA), which verifies credentials before a session is established. Older native RDP encryption was weaker and should not be relied on today.
Virtual channels are named data pipes multiplexed inside an RDP session, each carrying one type of traffic — clipboard, redirected drives and printers, audio, smart cards. They make the remote session feel integrated with your local machine.
Enforce TLS with Network Level Authentication, use strong unique passwords with an account lockout policy, keep the OS patched, restrict port 3389 by IP or put it behind a VPN/RD Gateway, and disable virtual-channel redirections you don't need.
