Transport Layer Services Overview

The transport layer ensures that the application processes on different hosts can communicate with each other as if they were directly connected. Think of it as a bridge between applications and the underlying network.

A transport-layer protocol, unlike network-layer protocols, is implemented within endpoints (hosts) and not in network routers.

On the sender side, the transport layer segments data into smaller chunks, adds a transport-layer header, and sends it to the network layer. The receiver reverses this process.

UDP (User Datagram Protocol) is a transport-layer protocol that offers an unreliable, connectionless service. This means data may be lost or arrive out of order.

TCP (Transmission Control Protocol) is a transport-layer protocol that delivers a reliable, connection-oriented service. It provides error detection and correction, ensuring data arrives in order.

The Internet Protocol (IP) operates at the network layer and is responsible for logical communication between hosts.

The transport layer provides a logical connection between applications, while the network layer provides a logical connection between hosts.

A measure of how much the SampleRTT (measured round-trip time) deviates from the EstimatedRTT (calculated round-trip time).

A formula used to calculate the DevRTT, which is a weighted average of the previous DevRTT value and the current difference between SampleRTT and EstimatedRTT. The value of β (beta) is typically set to 0.25, giving more weight to recent measurements. This ensures that the DevRTT value adapts to network changes more quickly.

The time interval that TCP uses to wait for an acknowledgment (ACK) before retransmitting a segment. It is calculated by adding 4 times the DevRTT to the EstimatedRTT. This formula helps to avoid premature timeouts by accounting for the potential variation in network conditions.

TCP uses a single retransmission timer to efficiently handle the retransmission of segments instead of maintaining individual timers for each segment. When a timeout occurs, TCP retransmits the segment associated with the timer and restarts the timer.

An acknowledgment segment (ACK) that acknowledges the receipt of multiple segments in sequence. The sendBase variable is updated to reflect the latest acknowledged byte, and the timer is restarted if there are remaining unacknowledged segments. This mechanism efficiently tracks data transmission progress and allows for quick responses to network congestion.

IP is a best-effort delivery service, meaning it tries to deliver segments but doesn't guarantee success. It doesn't ensure delivery, order, or data integrity, making it unreliable.

Each host on a network has a unique IP address used for identifying it and routing data.

UDP and TCP extend IP's host-to-host delivery to process-to-process delivery. They enable communication between applications running on different computers.

UDP provides process-to-process delivery and error checking, but doesn't ensure reliable data transfer, so it's considered unreliable.

TCP provides reliable data transfer using mechanisms like flow control, sequencing, acknowledgments, and timers, making it reliable.

TCP's congestion control mechanism prevents flooding the network with excessive traffic, ensuring fairness for all connections.

Demultiplexing is the process of delivering transport-layer segments to their corresponding sockets on the receiving host.

Multiplexing involves gathering data from different sockets on the sending host, packaging it into segments, and sending it to the network layer.

Each socket needs a unique identifier to ensure correct delivery of segments, achieved through source and destination port numbers.

A data transfer protocol employing error detection, positive acknowledgments, and negative acknowledgments.

The sender waits for the receiver to acknowledge receipt of a data packet before sending the next packet.

A packet that confirms the successful receipt of a data packet.

A packet that indicates that a data packet was received with errors and needs retransmission.

A sender sends a data packet and then waits for an ACK or NAK packet from the receiver.

A unique identifier assigned to each data packet sent to ensure proper sequencing and reliable delivery.

The process of sending a data packet again after it has been lost or corrupted.

Errors in ACK/NAK packets can lead to uncertainty about whether data has been received correctly.

Adding a unique sequence number to each data packet helps address the ambiguity of corrupted ACK/NAK packets.

The combination of error detection and sequence numbers ensures reliable data transfer.

Port numbers ranging from 0 to 1023 are reserved for well-known application protocols like HTTP (80) and FTP (21).

A unique identifier for a UDP socket, comprised of the destination IP address and destination port number. This allows multiple applications to use the same port as long as they have different source IPs or ports.

A unique identifier for a TCP socket encompassing the source IP address, source port number, destination IP address and destination port number.

The initial connection request sent from a TCP client to a server, containing details like destination port, source port, and a special connection-establishment bit in its header.

A TCP socket used by the server to listen for incoming connection establishment requests from clients on a specific port number (typically the server's default port).

A new TCP socket created on the server side specifically for a client, identifying the connection by the client's source port, IP address, and the server's port and IP address.

The process of data transfer between the client and server after a TCP connection has been established.

A server can handle multiple simultaneous TCP connections, each with its own dedicated socket and associated process.