TCP vs UDP at a Glance

At the transport layer, two protocols carry most internet traffic: TCP (Transmission Control Protocol) and UDP (User Datagram Protocol).

Both move data between applications, but they make very different trade-offs between reliability and speed. Knowing when each is used is core Network+ knowledge.

TCP is connection-oriented: before sending data it establishes a connection, ensuring both sides are ready. It guarantees that data arrives complete and in order.

If a segment is lost, TCP detects the gap and retransmits it. This reliability makes TCP the choice for web pages, email, and file transfers where every byte matters.

TCP starts a connection with a three-way handshake: SYN, SYN-ACK, ACK. The client sends SYN, the server replies SYN-ACK, and the client confirms with ACK.

Only after this exchange does data flow. The handshake synchronizes both sides and is a classic exam topic — remember the three steps in order.

UDP is connectionless: it sends data immediately with no handshake and no guarantee of delivery or order. It simply fires packets, called datagrams, toward the destination.

This makes UDP fast and lightweight, but the application must handle any lost or out-of-order data itself if it cares.

The core trade-off is clear: TCP adds reliability through acknowledgments and retransmission, which costs time and overhead. UDP skips all that for raw speed.

• TCP: reliable, ordered, slower.
• UDP: fast, lightweight, best-effort.

Choose based on whether accuracy or speed matters more.

Use TCP when every byte must arrive correctly. Web browsing (HTTP/HTTPS), email (SMTP), and file transfer (FTP) all rely on TCP.

A missing chunk of a web page or file would corrupt it, so the cost of TCP’s reliability is well worth it. Most internet traffic by volume is TCP.

Use UDP when speed and low latency beat perfect delivery. Live voice and video calls, online gaming, and streaming often use UDP.

In a video call, a slightly dropped frame is better than pausing to retransmit it. DNS lookups also use UDP for their quick, small queries.

TCP includes flow control to avoid overwhelming the receiver, using a sliding window that adjusts how much data is in flight before acknowledgment.

It also performs congestion control, slowing down when the network is busy. UDP has none of this, leaving rate control entirely to the application.

TCP’s header is larger (20 bytes minimum) because it carries sequence numbers, acknowledgments, and flags. UDP’s header is only 8 bytes, holding just source/destination ports and length.

The smaller UDP header means less overhead per packet, contributing to its speed advantage for high-volume real-time traffic.

To decide, ask: does the application need guaranteed, ordered delivery? If yes, TCP. If it values speed and can tolerate some loss, UDP.

Many protocols pick deliberately — DNS uses UDP for fast queries but switches to TCP for large transfers. Understanding the trade-off explains these design choices.

Both protocols use port numbers to direct traffic to the right application, the subject of the next lesson. They share the addressing scheme but differ in delivery guarantees.

Knowing TCP versus UDP is a foundation for understanding the specific protocols and ports that follow in this course.

Match the protocol to the need.

You compared TCP and UDP.

• TCP: connection-oriented, reliable, ordered; uses the SYN/SYN-ACK/ACK handshake.
• UDP: connectionless, fast, best-effort, no handshake.
• TCP suits web, email, and file transfer; UDP suits voice, video, gaming, and DNS queries.
• TCP adds flow and congestion control; UDP leaves that to the app.
• Both use port numbers to reach applications.