UDP and Connectionless Transport

The User Datagram Protocol is the Internet's lightweight connectionless transport, adding little more than process multiplexing and an optional checksum to IP's best-effort service, so applications gain minimal overhead and latency at the cost of reliability guarantees.

Definition

The User Datagram Protocol is a connectionless transport-layer protocol that provides best-effort, message-oriented delivery between processes, adding port-based multiplexing and an optional checksum to the network layer but offering no reliability, ordering, flow control, or congestion control.

Scope

This topic covers connectionless transport: the UDP segment format, its multiplexing and demultiplexing via port numbers, its optional checksum, and the reasons an application would choose it over TCP — no connection-setup delay, no congestion-control throttling, finer control over what and when to send, and small per-packet overhead. It surveys typical UDP-based applications and the responsibilities (such as reliability or rate control) that shift to the application. It excludes TCP's reliability and congestion machinery.

Core questions

• What minimal services does UDP add on top of IP?
• How does UDP multiplex and demultiplex data using port numbers?
• Why would an application prefer UDP to TCP?
• What responsibilities, such as reliability and rate control, move to the application with UDP?
• Which kinds of applications typically use UDP?

Key concepts

• connectionless service
• datagram (message) orientation
• UDP segment format
• port-based multiplexing/demultiplexing
• optional checksum
• low overhead and latency
• no congestion control
• application-layer reliability

Key theories

Minimal connectionless service

UDP exposes IP's best-effort datagram service to applications almost unaltered, adding only multiplexing by port and an optional checksum, so it has no setup, no state, and no automatic recovery — trading guarantees for simplicity and low latency.

Application-controlled transport

By avoiding TCP's congestion control and reliability, UDP gives an application direct control over timing and what to retransmit, which suits real-time and custom protocols, but obliges the application to handle loss and avoid overwhelming the network itself.

Clinical relevance

UDP underpins services where low latency or fine control matters more than guaranteed delivery: DNS lookups, real-time voice and video, online games, and network management. Newer transports such as QUIC are built on top of UDP precisely to gain control over the transport while bypassing TCP's constraints, making connectionless transport central to modern low-latency networking.

History

UDP was specified in RFC 768 (1980) as a minimal transport for applications that did not need TCP's machinery, when TCP and IP were separated into distinct protocols. It has remained essentially unchanged, and its role has grown as latency-sensitive applications proliferated and as developers built new transport features, including the QUIC protocol, on top of it.

Key figures

• Jon Postel
• James F. Kurose
• Keith W. Ross

Seminal works

• rfc768
• kurose2021

Frequently asked questions

If UDP is unreliable, why use it?

For many applications, fast delivery matters more than perfect delivery. Real-time voice and video can tolerate occasional loss but not delay, DNS benefits from a quick single-exchange query, and some applications implement their own reliability. UDP's lack of setup and congestion throttling gives them lower latency and more control.

Does UDP do anything beyond IP?

Yes, but little. UDP adds port numbers so that data reaches the correct application process on a host (multiplexing and demultiplexing), and an optional checksum to detect corruption. Beyond that it leaves IP's best-effort, connectionless delivery unchanged.