Transport Layer Complete Study Guide – Data Communication Chapter 3

What is the Transport Layer?

My dear students, welcome to Chapter 3! Today we are going to learn about the Transport Layer, which is the 4th layer in the OSI model. Now, let me ask you a simple question before we start. Have you ever wondered how, when you open your browser and also a messaging app at the same time, the right data goes to the right application? The transport layer does exactly that job.

The transport layer is responsible for process-to-process delivery of the entire message. Notice the words carefully — it says process-to-process, not just computer-to-computer. A process is simply an application program running on a host, like your web browser, your email client, or your messaging app.

Now, here is an important difference. The network layer (layer 3) handles source-to-destination delivery of individual packets. But here is the problem: the network layer does not recognize any relationship between those packets. It treats each packet independently, as if each piece belonged to a separate message. The transport layer, on the other hand, ensures that the whole message arrives intact and in order. It handles both error control and flow control at the source-to-destination level.

Major Functions of the Transport Layer

Let me list the major functions for you in a simple way:

1. It enables multiple applications to communicate over the network at the same time on a single device.

2. It ensures all the data is received reliably and in order by the correct application (if required).

3. It employs error handling mechanisms to detect and correct problems.

Five Responsibilities of the Transport Layer

Now, let us look at each responsibility one by one. These are very important for your exam, so please pay close attention.

1. Service-Point Addressing (Port Addressing)

Think about this: your computer runs several programs at the same time. You might have a browser open, an email client running, and maybe a file transfer application. When data arrives at your computer, how does the system know which program should receive which data?

The answer is service-point addressing. The transport layer header includes a type of address called a service-point address, which is also known as a port address.

The network layer gets each packet to the correct computer; the transport layer gets the entire message to the correct process on that computer. This is a very common exam sentence, so remember it well!

2. Segmentation and Reassembly

A message from the application layer can be very large. The transport layer divides the message into transmittable segments. Each segment is given a sequence number. These sequence numbers are very important because they allow the transport layer at the receiving end to reassemble the message correctly, even if the segments arrive out of order. They also help to identify and replace packets that were lost during transmission.

3. Flow Control

Flow control at the transport layer works similarly to flow control at the data link layer, but there is an important difference. Flow control at the transport layer is performed end to end, not across a single link. This means the sender and receiver adjust the rate of data transfer based on the receiver’s ability to handle data, even though the data may travel through many intermediate links.

4. Error Control

Just like flow control, error control at the transport layer is also different from the data link layer. Error control at the transport layer is performed process-to-process, not across a single link. The sending transport layer makes sure that the entire message arrives at the receiving transport layer without error — no damage, no loss, and no duplication. Error correction is usually achieved through retransmission.

5. Connection Control

The transport layer can operate in two modes:

A. Connectionless: Each segment is treated as an independent packet. There is no connection establishment before sending, and no connection release after. Think of it like sending a postcard — you just drop it in the mailbox without checking if the receiver is ready.

B. Connection-oriented: A connection is first established with the transport layer at the destination machine. Then data is transferred. After all data is transferred, the connection is terminated. Think of it like making a phone call — you first establish the call, talk, and then hang up.

Process-to-Process Delivery

Now let me explain this concept very clearly because students often get confused here. Let us revise the delivery types at each layer:

Real communication on the Internet is not between two nodes or two hosts. Real communication takes place between two processes (application programs). At any moment, several processes may be running on the source host and several on the destination host. The transport layer uses port numbers to deliver data from the correct process on the source to the correct process on the destination. Two processes communicate in a client/server relationship.

TCP Segment Format – Detailed Breakdown

A packet in TCP is called a segment. The segment consists of a 20 to 60 byte header, followed by data from the application. The header is 20 bytes if there are no options, and up to 60 bytes if options are included. Let me walk you through each field one by one.

Source Port Address (16 bits)

This field defines the port number of the application program on the host that is sending the segment. It tells the receiver which process sent the data.

Destination Port Address (16 bits)

This field defines the port number of the application program on the host that is receiving the segment. It tells the receiver which process should get the data.

Sequence Number (32 bits)

This is a very important field. It defines the number assigned to the first byte of data in this segment. TCP numbers every single byte that it transmits. The sequence number tells the destination which byte in the overall stream is the first byte in this segment. During connection establishment, each party uses a random number generator to create an Initial Sequence Number (ISN).

Acknowledgment Number (32 bits)

This field defines the byte number that the receiver is expecting to receive next. If the receiver has successfully received all bytes up to byte number x, it sets the acknowledgment number to x + 1. This is called cumulative acknowledgment. Acknowledgment and data can also be sent together in the same segment — this is called piggybacking.

Header Length (4 bits)

This field indicates the number of 4-byte words in the TCP header. Since the header can be between 20 and 60 bytes, the value of this field ranges from 5 (5 × 4 = 20 bytes) to 15 (15 × 4 = 60 bytes).

Control Flags (6 bits)

These are six individual bits, each with a specific purpose:

Window Size (16 bits)

This field defines the size of the sliding window, in bytes, that the other party must maintain. The maximum window size is 65,535 bytes (because 16 bits = 2^16 = 65,536 values). This value is called the receiving window (rwnd) and is determined by the receiver. The sender must follow this limit for flow control.

Checksum (16 bits)

This field contains the checksum for error detection. Unlike UDP where the checksum is optional, in TCP the checksum is mandatory. For the TCP pseudo header, the protocol field value is 6.

Urgent Pointer (16 bits)

This field is valid only when the URG flag is set. It defines a number that must be added to the sequence number to find the last urgent byte in the data section.

Options (0 to 40 bytes)

Optional information can be included in the TCP header, up to 40 bytes.

Solved Example: Reading a TCP Header from Hexadecimal

This type of question comes very often in exams. Let me teach you step by step how to solve it.

Transport Layer Addressing – Port Numbers

Now, let me explain port numbers very clearly. At every layer, we need an address to deliver data to the correct destination:

Port numbers are 16-bit integers, ranging from 0 to 65,535. The client program uses a randomly chosen port number (called an ephemeral port). The server process must use a specific, well-known port number so that clients can find it.

IANA Port Number Ranges

The Internet Assigned Numbers Authority (IANA) divides port numbers into three ranges:

Port Numbers in Client-Server Communication

Let me explain this with a simple real-life example. Imagine you go to a large office building (the host). You want to meet a specific person (the process). The building address is the IP address, and the person’s office room number is the port number.

When a client sends a message:

• The destination port is the well-known port of the server service (e.g., 80 for HTTP)
• The source port is a randomly generated ephemeral port to track this specific conversation
• This allows multiple simultaneous conversations — for example, many clients can request web pages from the same server at the same time, and each conversation is tracked by different source ports

Socket Addresses

A socket address is the combination of an IP address and a port number. Process-to-process delivery needs a pair of socket addresses:

Multiplexing and Demultiplexing

Now, class, let me ask you a question. What happens when multiple applications on your computer want to send data at the same time, but there is only one network connection? The transport layer solves this using multiplexing and demultiplexing.

Multiplexing (At the Sender)

At the sender side, several processes may need to send packets. But there is only one transport layer. This is a many-to-one relationship. The transport layer accepts messages from different processes (identified by their port numbers), adds headers, and passes them to the network layer through a single connection.

Demultiplexing (At the Receiver)

At the receiver side, the relationship is one-to-many. The transport layer receives datagrams from the network layer, removes the header, and delivers each message to the appropriate process based on the port number in the header.

Connectionless vs Connection-Oriented Services

Connectionless Service

In a connectionless service, packets are sent with no need for connection establishment or connection release. The packets are not numbered. They may be delayed, lost, or arrive out of sequence. There is no acknowledgment. UDP is connectionless.

Connection-Oriented Service

In a connection-oriented service, a connection is first established between sender and receiver. Then data is transferred. At the end, the connection is released. TCP is connection-oriented.

Reliable vs Unreliable Services

A reliable transport layer protocol implements flow and error control. It is slower but ensures all data arrives correctly. A unreliable protocol is faster but does not guarantee delivery. It is suitable for real-time applications.

Transport Layer Protocols

Simple Protocol

This is the most basic protocol — connectionless with no flow control and no error control. It assumes the receiver can immediately handle any packet. There is no acknowledgment, no sequencing, and no retransmission.

Stop-and-Wait Protocol

This is a connection-oriented protocol that uses both flow and error control. The sender and receiver both use a sliding window of size 1. Here is how it works:

1. The sender sends one packet at a time
2. The sender waits for an acknowledgment before sending the next one
3. A checksum is added to detect corrupted packets
4. The sender keeps a copy of the packet until the acknowledgment arrives

Go-Back-N (GBN) Protocol

The Go-Back-N protocol is more efficient than Stop-and-Wait. The key idea is that the sender can send several packets before receiving acknowledgments, but the receiver can only buffer one packet. Important points:

• The sender keeps copies of sent packets until acknowledgments arrive
• The receiver keeps track of only one variable (the next expected packet)
• There is no need to buffer out-of-order packets — they are simply discarded
• If a packet is lost or corrupted, the receiver discards all subsequent packets and the sender must go back and retransmit all packets from the lost one onward
• This protocol is inefficient if the network loses many packets

Selective-Repeat (SR) Protocol

The Selective-Repeat protocol solves the inefficiency of Go-Back-N. Instead of retransmitting everything, it resends only the packets that were actually lost. Important points:

• The receiver can buffer out-of-order packets
• Only the lost or corrupted packets are retransmitted
• The receiver forwards packets to the transport layer only when they arrive in order and acknowledgments arrive on time
• More efficient than Go-Back-N, but more complex at the receiver side

User Datagram Protocol (UDP) – Complete Details

UDP stands for User Datagram Protocol. It is called a connectionless, unreliable transport protocol. It does not add much to the services of IP except to provide process-to-process communication using port numbers. It performs very limited error checking.

Now you might wonder: if UDP is so basic, why would anyone use it? Here is the answer: UDP has very low overhead, which makes it fast. If a process wants to send a small message and does not care much about reliability, UDP is a great choice.

UDP Header Format

UDP packets are called user datagrams. They have a fixed-size header of 8 bytes:

Source Port (16 bits): Port number of the sending process.

Destination Port (16 bits): Port number of the receiving process.

Length (16 bits): Total length of the user datagram (header + data).

Checksum (16 bits): Used for error detection. Optional in UDP!

What UDP Does NOT Do

• No flow control
• No error control (beyond optional checksum)
• No connection control
• No segmentation and reassembly
• No sequence numbers

All these functions must be handled by the application layer programs if needed. Because UDP cannot segment and reassemble, it can only transmit small portions of data that fit into a single datagram.

What UDP DOES Do

• Service-point addressing (using port numbers)
• Process-to-process delivery

When is UDP Used?

1. Simple request-response communication — when you need a quick exchange with little concern for reliability
2. Applications with internal flow and error control — for example, TFTP (Trivial File Transfer Protocol) has its own error control
3. Multicasting — UDP supports multicasting, TCP does not
4. Management protocols — like SNMP (Simple Network Management Protocol)
5. Route updating protocols — like RIP (Routing Information Protocol)

Solved Example: Reading a UDP Header from Hexadecimal

TCP (Transmission Control Protocol) – Complete Details

Now we come to the most important protocol in the transport layer. TCP is a connection-oriented, reliable transport protocol. Unlike UDP, TCP provides many services that ensure data is delivered correctly and in order.

TCP Services

1. Process-to-Process Communication — using port numbers, just like UDP
2. Stream Delivery Service — data is sent as a continuous stream of bytes
3. Full-Duplex Communication — data can flow in both directions at the same time
4. Connection-Oriented Service — connection is established before data transfer
5. Reliable Service — error control and flow control ensure reliable delivery

Stream Delivery Service

This is a very important concept. In UDP, a process sends messages with predefined boundaries. Each message becomes one datagram. But in TCP, there are no boundaries. TCP allows the sending process to deliver data as a stream of bytes, and the receiving process reads it as a stream of bytes. Think of it like water flowing through a pipe — there is no concept of where one “piece” of water ends and the next begins.

Sending and Receiving Buffers

Because the sending and receiving processes may not write or read data at the same speed, TCP uses buffers for storage. There are two buffers: a sending buffer and a receiving buffer, one for each direction. These are usually implemented as circular arrays of byte locations. The buffers can be hundreds or thousands of bytes in size.

How TCP Creates Segments

The sending process writes bytes into the sending buffer. TCP then groups a number of bytes together, adds a header, and creates a segment. The segment is then given to the IP layer for transmission. All of this is transparent to the receiving process.

TCP Segment Numbering System

Byte Numbering: TCP numbers every single byte in a connection. Numbering is independent in each direction. The numbering does not start from 0. Instead, TCP generates a random number between 0 and 2^32 – 1 as the number of the first byte. This random starting number is called the Initial Sequence Number (ISN).

Sequence Number: The sequence number of a segment is simply the byte number of the first byte carried in that segment.

Acknowledgment Number

Since TCP is full-duplex, both parties send and receive data at the same time. Each party uses an acknowledgment number to confirm the bytes it has received. The acknowledgment number defines the number of the next byte expected. It is cumulative, meaning if a party uses 5643 as an acknowledgment number, it has received all bytes up to 5642.

Full-Duplex Communication

TCP allows data to flow in both directions at the same time. Each TCP connection has a sending buffer and a receiving buffer, and segments move in both directions simultaneously.

Connection-Oriented Service

TCP creates a virtual connection (not physical). The process is: establish connection → exchange data → terminate connection. Even though the segments travel through IP datagrams that may take different paths, TCP creates the illusion of a dedicated connection.

Flow Control in TCP

The receiver controls the amount of data the sender can transmit. This prevents the receiver from being overwhelmed. TCP uses a byte-oriented flow control with the window size field in the header. The maximum window size is 65,535 bytes.

Error Control in TCP

TCP provides reliability through error control. It handles:

• Corrupted segments — detected using checksum
• Lost segments — detected using timeouts and acknowledgments
• Out-of-order segments — stored temporarily but not delivered to the process until in order
• Duplicated segments — detected using sequence numbers

Error detection and correction uses three tools:

1. Checksum — mandatory 16-bit checksum in every segment. If corrupted, the segment is discarded and treated as lost.
2. Acknowledgment — confirms receipt of data segments. ACK segments are never acknowledged themselves.
3. Timeout — if an acknowledgment is not received within a certain time, the segment is retransmitted.

Congestion Control

TCP also considers congestion in the network. The amount of data sent is not only controlled by the receiver (flow control) but also by the level of congestion in the network. If the network is congested, TCP reduces its sending rate.

Quick Comparison: UDP vs TCP

Challenge Exam Questions – Transport Layer

Dear students, these questions are designed to test your deep understanding. Try each one before checking the answer!