Networks: Unit 7 - TCP and UDP

What is the main purpose of the transport layer?

To provide logical communication between processes running on different hosts.

What does 'logical communication' mean in the transport layer?

Applications behave as if hosts are directly connected even though the network may be complex.

Where is the transport layer implemented?

In end systems (hosts), not in network routers.

What is a transport segment?

Application data plus a transport layer header.

What are the two main Internet transport protocols?

TCP and UDP.

What are key characteristics of TCP?

Reliable, connection-oriented, flow-controlled, and congestion-controlled.

What are key characteristics of UDP?

Unreliable and connectionless.

What services are NOT provided by the transport layer?

Delay guarantees and bandwidth guarantees.

What is multiplexing?

Combining data from multiple processes and attaching headers before sending.

What is demultiplexing?

Delivering received segments to the correct application socket using header information.

How is a TCP socket identified?

By a four-tuple: source IP, source port, destination IP, destination port.

How is a UDP socket identified?

By destination IP and destination port only.

Why do some applications prefer UDP?

It has lower delay, no handshake, and low overhead.

Give examples of applications that use TCP.

HTTP, FTP, SMTP, Telnet.

Give examples of applications that use UDP.

DNS, VoIP, streaming, SNMP.

What is the purpose of the UDP checksum?

Basic error detection.

What is the goal of reliable data transfer (rdt)?

To deliver data without loss, corruption, or reordering.

What is an ACK?

Acknowledgment that a packet was received correctly.

What is a NAK?

A signal indicating a packet was received with errors.

How does the receiver handle duplicate packets?

Discards duplicates and resends ACK.

What is a limitation of stop-and-wait protocols?

Only one packet can be in flight → low utilization.

What is pipelining?

Sending multiple packets without waiting for ACKs.

Why is pipelining beneficial?

Greatly increases throughput.

What are the two pipelined protocols?

Go-Back-N (GBN) and Selective Repeat (SR).

What is the window size in pipelined protocols?

Number of unacknowledged packets allowed in flight.

Why do pipelining protocols need larger sequence numbers?

To uniquely identify all packets in flight.

How does the Go-Back-N (GBN) receiver handle out-of-order packets?

Discards them and sends cumulative ACK.

When does Go-Back-N (GBN) retransmit packets?

When the oldest unacknowledged packet times out.

How does the Selective Repeat (SR)receiver handle out-of-order packets?

Buffers them and ACKs individually.

What is a benefit of Selective Repeat (SR) over Go-Back-N (GBN) ?

Only erroneous packets are resent, conserving bandwidth.

What type of service does TCP provide?

Reliable, full-duplex byte-stream service.

What is Maximum Segment Size (MSS) in TCP?

Maximum Segment Size—max data bytes in a TCP segment.

What does a TCP sequence number represent?

The byte number of the first byte in the segment.

What does a TCP acknowledgment number represent?

The next byte expected by the receiver.

When is the TCP ACK flag set?

In all segments except the initial SYN.

What are common TCP header fields?

Ports, sequence number, acknowledgment number, flags, window, checksum, options.

What are the TCP control flags?

SYN, ACK, FIN, RST, URG, PSH.

What is the TCP three-way handshake used for?

Establishing a connection and agreeing on sequence numbers.

What happens in the first step of the three-way handshake?

Client sends SYN with its initial sequence number.

What happens in the second handshake step?

Server responds with SYN-ACK.

What happens in the final handshake step?

Client sends ACK confirming the connection.

What does 'full duplex' mean in TCP?

Data can flow simultaneously in both directions.

What is flow control in TCP?

Prevents sender from overwhelming the receiver.

What does the TCP receiver window represent?

The number of bytes the receiver is willing to accept.

What is the purpose of the urgent pointer?

Indicates the location of urgent data in the segment.

Why does TCP include a checksum?

For error detection.

What is network congestion?

Too much sending too fast for the network to handle.

What are signs of congestion?

Packet loss and long queueing delays.

What does AIMD stand for?

Additive Increase, Multiplicative Decrease.

How does additive increase work?

cwnd grows by 1 MSS per RTT.

How does additive increase work?

cwnd grows by 1 MSS per RTT.

How does multiplicative decrease work?

cwnd is halved after packet loss.

What is the congestion window (cwnd)?

Controls how much data TCP can have in flight.

What is slow start used for?

Quickly ramping up the sending rate at the beginning of a connection.

What is the initial cwnd in slow start?

1 MSS.

When does slow start end?

When cwnd reaches threshold or a loss occurs.

What indicates packet loss to TCP?

Timeout or 3 duplicate ACKs.

How is a TCP connection closed?

Using a four-segment FIN/ACK exchange.

What port does Telnet use?

TCP port 23.

What port does HTTP use?

TCP port 80.

What port does DNS use?

UDP port 53.

What is the main difference between TCP and UDP?

TCP is reliable and connection-oriented; UDP is unreliable and connectionless.

Why is buffering required in pipelined protocols?

Multiple packets are in flight and must be tracked.

Why does UDP not use congestion control?

It is designed for real-time applications requiring steady transmission.

What ensures ordered delivery in TCP?

Sequence numbers and acknowledgment flow.

Why is the ACK number critical in TCP?

It ensures reliable and in-order byte delivery.