Computer Networks - 4

In layered architecture, the active elements in each layer are formally called entities. These entities perform the functions of that layer and provide services to the layer above.

When entities in layer n implement services used by layer n+1, layer n is called the service provider. The provider offers services that the layer above consumes.

The layer that consumes the service provided by the layer below it is known as the service user.

A Service Access Point (SAP) is the specific location or interface where a layer makes its services available to the layer above it.

SAP stands for Service Access Point, which is a conceptual location where one layer provides services to another.

PDU stands for Protocol Data Unit, which is the unit of data exchanged between peer entities in the same layer.

SDU stands for Service Data Unit, which is the data unit passed between layers when requesting services. It's the user data that needs to be transmitted.

Connectionless service is like a postal system where each message is independent and carries full destination address. No prior connection setup is needed; each message is treated separately.

Connection-oriented service is like a telephone system where a connection must be established before communication. The connection is maintained throughout the conversation.

In connectionless service, each message must contain the full destination address since there's no pre-established connection path. Every message is routed independently.

Reliable services are designed to ensure that data is delivered correctly and completely; therefore, they are defined by the fact that they do not lose data during transmission.

Message-oriented services are typically connection-oriented, ensuring that messages are delivered reliably and in order. Connectionless services, by contrast, do not guarantee delivery or order.

In a message-oriented service, the boundaries between individual messages are preserved, meaning the receiver gets exactly what the sender sent as a discrete unit. This contrasts with byte streams, where boundaries are not maintained.

A byte stream service, such as TCP, is connection-oriented and provides a reliable delivery mechanism for data. It ensures that bytes arrive in the correct order without loss.

In a byte stream, the data is treated as a continuous flow of bytes without any inherent structure or message boundaries. The application layer is responsible for interpreting the data stream if boundaries are needed.

Datagram services, such as UDP, are connectionless and do not provide reliability guarantees like error correction or retransmission. They are designed for speed rather than guaranteed delivery.

A sequence of pages is best represented as a message stream because each page is a distinct unit of information that needs to be handled as a whole. This preserves the logical structure of the document.

Remote login (like SSH or Telnet) requires a continuous, reliable flow of characters between the client and the server. This is the classic use case for a byte stream.

Electronic junk mail (email) is typically sent as individual, independent messages. Datagrams are suitable for this as they are discrete units of data that do not require a persistent connection for each individual packet.

Digitized voice (VoIP) is often transmitted using unreliable services (like UDP) because real-time delivery is more important than perfect reliability. Dropped packets are preferable to the latency caused by retransmission.

Registered mail implies a reliable, ordered, and connection-oriented delivery process, which is analogous to a byte stream or reliable connection-oriented service in networking contexts.

A database query involves sending a request and waiting for a specific reply. This request-reply pattern is a distinct communication model used in client-server interactions.