Principles of layering

Why networks use layers

Definition. A layered architecture divides communication into modules with defined services and interfaces. A layer uses services from the layer below and offers services to the layer above. This limits how much one protocol must know about other protocols. Example. An application can send a message without knowing which radio channel or copper cable carries its bits. Exam point. A layer boundary is an abstraction boundary; it does not imply a separate physical box.

Service, interface, and protocol

Definition. A service describes what a layer provides to its users. An interface specifies how the upper layer requests that service. A protocol is the set of rules and message formats used by peer entities to provide the service across the network. Example. A transport service may offer reliable byte delivery; its interface exposes send and receive calls; TCP is a protocol that implements such behavior. Exam point. Service describes the user-visible result, while protocol describes peer interaction.

Encapsulation and decapsulation

Definition. Each sender layer adds control information around or alongside data from the layer above. The receiver removes the corresponding information in reverse order. A layer’s payload is commonly called the service data unit, while its complete protocol unit includes its own control information. Example. An application message becomes a transport segment, then an IP datagram, then a link-layer frame, and finally a bit sequence on a medium. Exam point. Headers are added per encapsulating layer; a link-layer trailer such as an FCS may also be added.

Stage	Typical unit	Added information
Application	message	application syntax or request fields
Transport	segment/datagram	ports, sequence state, checksum
Network	packet/datagram	source and destination IP addresses
Link	frame	local addresses, type, error check
Physical	bits/symbols	encoding and signal representation

OSI reference model

Definition. The OSI model names seven conceptual layers: physical, data link, network, transport, session, presentation, and application. It is a reference framework, not a requirement that every deployed protocol implement seven separate modules. Its names help classify functions. Example. Routing is a network-layer function, while framing and medium access are data-link functions. Exam point. Do not confuse the OSI reference model with the TCP/IP protocol suite.

Layer, bottom-up	Typical responsibility
Physical	signals, bits, connectors, modulation
Data link	local frame delivery, MAC, link error detection
Network	logical addressing and routing across networks
Transport	process-to-process delivery, reliability, flow control
Session	dialogue organization and checkpoints in the model
Presentation	data representation, transformation, encryption concepts
Application	network application protocols and services

TCP/IP model and layer mapping

Definition. A common Internet model groups functions into link, internet, transport, and application layers. Some descriptions split physical from link or split the application layer into OSI session/presentation/application. Mapping is approximate because models organize functions differently. Example. IP belongs to the internet layer; TCP and UDP belong to transport; HTTP and DNS belong to application. Exam point. IP routing is not a transport-layer function even though transport data is carried inside IP.

TCP/IP layer	Common protocols or functions	Rough OSI correspondence
Application	HTTP, DNS, SMTP	application, presentation, session
Transport	TCP, UDP	transport
Internet	IPv4, IPv6, ICMP	network
Link	Ethernet, Wi-Fi, framing	data link and physical