--- title: "Why two transport protocols exist" subject: "Computer Networks and Security" chapter: 5 key: "semester_7/computer_networks_and_security:why-two-transport-protocols" source: https://www.bctengineer.com/q/semester_7/computer_networks_and_security/why-two-transport-protocols site: BCT Engineering --- # Why two transport protocols exist Computer Networks and Security · Chapter 5: Transport Layer - Asked in: 2069 Chaitra, 5 marks - Answer checked by Claude Opus 5.5 (high effort) on 2026-10-14 - Page: https://www.bctengineer.com/q/semester_7/computer_networks_and_security/why-two-transport-protocols - Whole chapter: https://www.bctengineer.com/md/semester_7/computer_networks_and_security/chapter-5 ## Question Why do you think there exist two protocols in the transport layer whereas there exists only one protocol in the Internet layer in the TCP/IP reference model? ## Short answer IP provides uniform best-effort packet delivery, so only one protocol is needed. Applications require different services: TCP offers reliable, ordered delivery for web/email, while UDP provides low-latency, lightweight transport for real-time apps like VoIP and DNS. This diversity necessitates two transport protocols. ## Answer The Internet layer has one protocol (IP) because its job is the same for everyone: deliver a packet from one host to another host using best-effort. The applications above it, however, need very different services, which one transport protocol cannot offer efficiently. So the transport layer has two main protocols, TCP and UDP. | Need of application | Protocol | |---|---| | Reliable, ordered, error-free delivery (web, email, file transfer) | TCP | | Speed, low delay, small messages, tolerance of loss (DNS query, VoIP, video, games) | UDP | #### Reasons 1. **TCP's reliability costs time.** Handshake, acknowledgements, retransmission and congestion control add delay and header overhead (20 to 60 bytes). Real-time applications prefer to lose a packet than to wait for it. 2. **UDP is light.** An 8-byte header, no connection set-up, no state in the end hosts; suitable for a single request-response such as DNS. 3. **Multicast and broadcast** can be done by UDP only, since TCP is point to point. 4. **Flexibility.** Applications that need their own reliability rules can build them over UDP (e.g. QUIC). 5. **IP needs only one protocol** because it only handles addressing and routing; every packet needs the same service at that layer, so no alternative protocol is needed. Both TCP and UDP run on top of it (IP protocol numbers 6 and 17). ## Check yourself The answer's key points as multiple-choice questions, with the correct option and why. 1. Why does the Internet layer need only one protocol? A. Its job (best-effort host-to-host delivery) is the same for all applications B. It guarantees reliability for everyone C. UDP and TCP both replace it D. It uses port numbers Answer: A. IP only handles addressing and routing, which is common to all. (Key point: Why IP is single) 2. Why do real-time applications often prefer UDP over TCP? A. UDP is encrypted B. TCP reliability adds delay and overhead; losing a packet is better than waiting C. UDP has larger headers D. UDP guarantees order Answer: B. Handshake, acknowledgements and retransmission add delay that real-time apps avoid. (Key point: TCP cost) 3. Which is true of UDP but not TCP? A. It supports multicast and broadcast B. It gives ordered delivery C. It uses a 20 to 60 byte header D. It has congestion control Answer: A. TCP is point to point; only UDP can do multicast and broadcast. (Key point: UDP features) ## Related in Computer Networks and Security - [Compare OSI and TCP/IP models](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Aosi-vs-tcpip) — Computer Networks and Security, chapter 1, asked 4× - [Why IP address when MAC exists](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Amac-vs-ip-address) — Computer Networks and Security, chapter 4 - [Functions of physical layer](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Aphysical-layer-functions) — Computer Networks and Security, chapter 2 - [Why network layer is key layer](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Anetwork-layer-key-layer) — Computer Networks and Security, chapter 4 - [Need of classless IP addressing](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Aclassless-need-example) — Computer Networks and Security, chapter 4 - [TCP header and its fields](https://www.bctengineer.com/question.md?key=semester_7%2Fcomputer_networks_and_security%3Atcp-header-structure) — Computer Networks and Security, chapter 5, asked 2× ## Related in other subjects - [TCP/IP applications; IPv4 vs IPv6 routing](https://www.bctengineer.com/question.md?key=semester_8%2Finternet_and_intranet%3Atcpip-application-ipv4-ipv6-routing) — Internet and Intranet, chapter 2 - [Intranet vs Internet and intranet protocols](https://www.bctengineer.com/question.md?key=semester_8%2Finternet_and_intranet%3Aintranet-vs-internet-protocols) — Internet and Intranet, chapter 1 - [Why DTFS uses only one period](https://www.bctengineer.com/question.md?key=semester_4%2Fsignals_and_systems%3Actfs-vs-dtfs-duration) — Signals and Systems, chapter 2 - [Isochrones, one and two way drainage](https://www.bctengineer.com/question.md?key=semester_4%2Fsoil_mechanics%3Aisochrones-one-two-way-drainage) — Soil Mechanics, chapter 9