Computer Networking Fundamentals: Protocols, Layers, and Architecture

Computer networking is the technical infrastructure that makes distributed computing possible, governing how devices exchange data across local, regional, and global scales. This page covers the formal layer model that structures networking protocols, the mechanisms by which data moves from one host to another, the scenarios in which different protocol families are applied, and the boundaries that distinguish one architectural choice from another. The discipline is standardized through bodies including the Internet Engineering Task Force (IETF) and the Institute of Electrical and Electronics Engineers (IEEE), whose specifications define interoperability across virtually all modern networks.


Definition and scope

Computer networking encompasses the hardware, software, and protocol systems that enable two or more computing devices to exchange data. The scope spans physical transmission media (copper, fiber, radio spectrum), the logical rules governing data formatting and routing, and the application-level conventions that programs use to communicate. As a subdomain of computer science, networking is inseparable from distributed systems, operating systems fundamentals, and cybersecurity fundamentals.

The canonical reference model for networking is the OSI (Open Systems Interconnection) model, published by the International Organization for Standardization (ISO) as ISO/IEC 7498-1. The OSI model divides network communication into 7 discrete layers:

  1. Physical — raw bit transmission over a medium (e.g., Ethernet cable, fiber optic)
  2. Data Link — framing, MAC addressing, and error detection across a single link (IEEE 802.x standards)
  3. Network — logical addressing and routing across multiple hops (IP)
  4. Transport — end-to-end delivery, segmentation, and flow control (TCP, UDP)
  5. Session — establishment, management, and termination of communication sessions
  6. Presentation — data translation, compression, and encryption formatting
  7. Application — protocols used directly by applications (HTTP, DNS, SMTP)

In practice, the TCP/IP model — defined across a suite of IETF Request for Comments (RFC) documents — collapses the OSI model into 4 layers: Link, Internet, Transport, and Application. TCP/IP is the operational architecture underlying the public Internet. The IETF publishes and maintains these specifications at ietf.org.


How it works

Data transmission in a packet-switched network follows a process called encapsulation. When an application sends a message, each layer of the networking stack adds its own header (and sometimes a trailer) to the payload before passing it downward toward the physical medium. The receiving device reverses this process — decapsulation — stripping headers as data moves up through the stack.

The core delivery mechanism depends on the Transport layer protocol selected:

At the Network layer, IP (Internet Protocol) handles logical addressing. IPv4, defined in IETF RFC 791, uses 32-bit addresses supporting approximately 4.3 billion unique addresses. IPv6, defined in IETF RFC 8200, uses 128-bit addresses, yielding 3.4 × 10³⁸ possible addresses — a pool designed to eliminate address exhaustion permanently.

Routing between networks depends on protocols that fall into two categories:


Common scenarios

Local Area Networks (LANs) connect devices within a single building or campus. IEEE 802.3 (Ethernet) governs wired LANs; IEEE 802.11 (Wi-Fi) governs wireless LANs. The IEEE 802.11ax standard (Wi-Fi 6) supports theoretical throughput of up to 9.6 Gbps per access point, distributed across connected clients (IEEE 802.11ax).

Wide Area Networks (WANs) interconnect geographically separated LANs. The public Internet is the largest WAN. Enterprise WANs commonly use MPLS (Multiprotocol Label Switching) for traffic engineering, or SD-WAN architectures that virtualize WAN connectivity over broadband circuits.

Content Delivery Networks (CDNs) distribute cached content across geographically dispersed servers to reduce latency. DNS-based routing directs client requests to the nearest available edge node, a technique formalized through Anycast addressing (IETF RFC 4786).

IoT networks apply constrained-environment protocols, including MQTT (defined in the OASIS MQTT 5.0 specification) and CoAP (IETF RFC 7252), which operate with message overhead as low as 4 bytes per CoAP header — critical for battery-powered devices. This overlaps directly with the scope of internet of things architectures.


Decision boundaries

Choosing among protocols and architectures requires evaluating four axes: reliability requirements, latency sensitivity, scalability targets, and security posture.

TCP vs. UDP is the primary transport-layer decision. TCP is appropriate for file transfer (FTP, HTTP), email (SMTP), and any transaction requiring guaranteed delivery. UDP is appropriate for real-time audio/video streaming, DNS queries (where retransmission logic is handled at the application layer), and online gaming. HTTP/3, standardized in IETF RFC 9114, builds on QUIC — a UDP-based transport — to achieve TCP-equivalent reliability without head-of-line blocking.

IPv4 vs. IPv6 is largely determined by infrastructure context. Legacy enterprise environments may remain on IPv4 with NAT (Network Address Translation) for address conservation; public-facing services and government networks are subject to IPv6 adoption mandates. The U.S. Office of Management and Budget issued OMB Memorandum M-21-07 requiring federal agencies to achieve at least 80% IPv6-only operation on federal networks by fiscal year 2025.

Centralized vs. distributed routing affects resilience. A single OSPF domain offers simpler management within an autonomous system but creates a single failure domain. BGP-based architectures distribute routing decisions across autonomous systems, enabling Internet-scale fault tolerance at the cost of configuration complexity and convergence time measured in seconds to minutes.

Network security intersects directly with architectural decisions — a topic covered in depth at network security principles and cryptography in computer science.


References