TCP 및 UDP 기초

When data travels across a network, it uses different rules or 'protocols'. At the Transport layer of the TCP/IP model, two main protocols handle how applications send and receive data: TCP and UDP.

They both move data, but they do it in fundamentally different ways, each suited for specific tasks.

TCP stands for Transmission Control Protocol. Think of TCP as a very careful post office that guarantees your letter will arrive, in order, and without errors.

• It's connection-oriented: A connection must be established before data is sent.
• It's reliable: It guarantees delivery of data.
• It ensures ordered data transfer and error checking.

Before TCP sends any application data, it performs a 'three-way handshake' to establish a connection:

1. SYN (Synchronize): Client asks to connect.
2. SYN-ACK (Synchronize-Acknowledge): Server acknowledges and agrees.
3. ACK (Acknowledge): Client acknowledges, and the connection is ready.

This handshake ensures both sides are ready to communicate reliably.

TCP uses several mechanisms to ensure reliability:

• Acknowledgements (ACKs): The receiver sends ACKs for data received. If no ACK, the sender retransmits.
• Sequence Numbers: Data packets are numbered to ensure they arrive in the correct order and to detect missing packets.
• Flow Control: Prevents a fast sender from overwhelming a slow receiver.
• Congestion Control: Manages network traffic to avoid overloading the network.

This Python code snippet shows how a TCP client socket is created and attempts to connect. Notice the SOCK_STREAM type, which signifies TCP.

Run it to see the connection attempt!

UDP stands for User Datagram Protocol. Unlike TCP, UDP is like sending a postcard: you send it, and you hope it arrives, but there's no guarantee or tracking.

• It's connectionless: No connection is established before sending.
• It's unreliable: No guarantees of delivery, order, or error-free transmission.
• It prioritizes speed and low overhead over reliability.

The simplicity of UDP comes from its lack of features:

• There's no three-way handshake to set up a connection.
• There are no acknowledgements (ACKs) for received data.
• There's no retransmission of lost packets.
• There's no built-in flow or congestion control.

This means less overhead, making it much faster for certain applications.

This Python code demonstrates a UDP client sending data. Notice the SOCK_DGRAM type, which indicates UDP. Because it's connectionless, it can send data directly.

Run it to see a UDP datagram being sent!

Here's a quick comparison of the two:

• Connection: TCP is connection-oriented; UDP is connectionless.
• Reliability: TCP guarantees delivery; UDP offers best-effort delivery.
• Order: TCP ensures ordered data; UDP does not.
• Speed: UDP is generally faster due to less overhead; TCP is slower due to reliability mechanisms.
• Overhead: TCP has higher overhead; UDP has lower overhead.

Choosing between TCP and UDP depends on your application's needs:

• Use TCP for: Web browsing (HTTP/HTTPS), Email (SMTP, IMAP, POP3), File Transfer (FTP), Secure Shell (SSH). When data integrity and order are critical.
• Use UDP for: Online gaming, Video/Audio streaming, Voice over IP (VoIP), Domain Name System (DNS). When speed and low latency are more important than guaranteed delivery.

You've explored the fundamental differences between TCP and UDP!

• TCP is reliable, connection-oriented, and ensures data integrity and order, but with higher overhead.
• UDP is fast, connectionless, and has low overhead, but offers no delivery guarantees.

Understanding these distinctions is crucial for designing and troubleshooting network applications. Next, you'll apply these concepts to real-world scenarios!