2.1 Network Models & Physical Layer

Key Takeaways

  • The 5-layer TCP/IP model simplifies the 7-layer OSI model by consolidating the Session, Presentation, and Application layers into a unified Application layer that directly aligns with modern Internet protocols.
  • Protocol Data Units (PDUs) change names and structure through the encapsulation process: Application Data becomes Transport Segments, Network Packets/Datagrams, Data Link Frames, and Physical Bits.
  • Ethernet copper cabling specifications dictate performance: Cat5e supports 1 Gbps up to 100 meters, Cat6 supports 10 Gbps up to 55 meters (1 Gbps at 100 m), and Cat6a supports 10 Gbps up to a full 100 meters using T568A or T568B termination pinouts.
  • Network devices operate at distinct layers: Layer 1 Hubs repeat signals across a single shared collision domain, Layer 2 Switches use MAC address tables (CAM tables) to microsegment collision domains per port, and Layer 3 Routers segment broadcast domains using IP routing tables.
Last updated: August 2026

Network Models & Physical Layer Architecture

Computer networking is the practice of interconnecting computing devices to exchange data and share resources using established communication protocols. To design, troubleshoot, and secure modern networks, IT professionals rely on layered abstraction models. These models divide the complex process of data communication into modular, manageable layers where each layer provides specific services to the layer above it and relies on the layer beneath it.


1. Reference Models: 5-Layer TCP/IP vs. 7-Layer OSI

Two primary architectural frameworks govern network communication theory and practice: the theoretical Open Systems Interconnection (OSI) 7-Layer Model created by the International Organization for Standardization (ISO), and the pragmatic TCP/IP 5-Layer Model (also known as the Internet Reference Model) that powers the modern Internet.

+-----------------------------------------------------------------------------------------+
|                         OSI 7-LAYER VS. TCP/IP 5-LAYER MODEL                            |
|                                                                                         |
|   OSI 7-Layer Theoretical Model                      TCP/IP 5-Layer Practical Model     |
|   +---------------------------------------+          +------------------------------+   |
|   | Layer 7: Application (HTTP, DNS, SSH) |          |                              |   |
|   +---------------------------------------+          |                              |   |
|   | Layer 6: Presentation (SSL/TLS, ASCII)| ------>  | Layer 5: Application         |   |
|   +---------------------------------------+          | (HTTP, HTTPS, DNS, DHCP, SSH)|   |
|   | Layer 5: Session (RPC, NetBIOS, Sockets)         |                              |   |
|   +---------------------------------------+          +------------------------------+   |
|   | Layer 4: Transport (TCP, UDP)         | ------>  | Layer 4: Transport (TCP, UDP)|   |
|   +---------------------------------------+          +------------------------------+   |
|   | Layer 3: Network (IPv4, IPv6, ICMP)   | ------>  | Layer 3: Network (IP, ICMP)  |   |
|   +---------------------------------------+          +------------------------------+   |
|   | Layer 2: Data Link (Ethernet, Wi-Fi)  | ------>  | Layer 2: Data Link (MAC)     |   |
|   +---------------------------------------+          +------------------------------+   |
|   | Layer 1: Physical (Cables, Radio, Bits| ------>  | Layer 1: Physical (Bits/PHY) |   |
|   +---------------------------------------+          +------------------------------+   |
+-----------------------------------------------------------------------------------------+

The OSI 7-Layer Model

The OSI model provides an exhaustive conceptual framework for network communications. A common mnemonic to remember the layers from 1 to 7 is "Please Do Not Throw Sausage Pizza Away" (Physical, Data Link, Network, Transport, Session, Presentation, Application):

  1. Layer 1 - Physical: Transmits unstructured raw bit streams over physical transmission media (electrical voltages, light pulses, or radio frequencies).
  2. Layer 2 - Data Link: Packages raw bits into structured Frames, handles physical hardware addressing via Media Access Control (MAC) addresses, and performs basic error detection using the Frame Check Sequence (FCS).
  3. Layer 3 - Network: Manages logical addressing (IP addresses), path determination, and forwarding of Packets across disparate networks via routing.
  4. Layer 4 - Transport: Manages end-to-end communication, session multiplexing via port numbers, segmentation, flow control, and reliability (e.g., TCP segment retransmission vs. UDP datagram delivery).
  5. Layer 5 - Session: Establishes, manages, maintains, and terminates communication sessions and dialogues between local and remote applications (e.g., RPC, NetBIOS).
  6. Layer 6 - Presentation: Handles syntax negotiation, data formatting, character encoding (ASCII, UTF-8), data compression, and cryptographic encryption/decryption (SSL/TLS).
  7. Layer 7 - Application: Provides network services directly to end-user software applications and system processes (e.g., web browsers, email clients).

The TCP/IP 5-Layer Model

While the OSI model is ideal for academic study and protocol troubleshooting, real-world protocol implementations consolidate Layers 5, 6, and 7 into a single unified Application Layer. In the TCP/IP stack, any application (such as an Apache web server or OpenSSH daemon) natively handles its own session state, encryption (TLS), and application-level payload formatting.

Layer Mapping & Protocol Data Units (PDUs)

As data moves through the protocol stack, each layer encapsulates the payload with a specific header (and sometimes a trailer), creating a distinct Protocol Data Unit (PDU).

Layer NumberOSI Layer NameTCP/IP Layer NameProtocol Data Unit (PDU)Core Addressing / IdentifiersPrimary Protocols & Standards
Layer 7 / 5ApplicationApplicationData / PayloadURLs, FQDNs, User IdentifiersHTTP, HTTPS, DNS, DHCP, SSH, SMTP, FTP
Layer 6Presentation(Consolidated in App)DataEncoding / MIME TypesTLS/SSL, JPEG, ASCII, JSON
Layer 5Session(Consolidated in App)DataSession IDs, SocketsRPC, NetBIOS, PPTP
Layer 4TransportTransportSegment (TCP) / Datagram (UDP)Port Numbers (0–65535)TCP, UDP, SCTP, QUIC
Layer 3NetworkNetwork (Internet)Packet / DatagramLogical IP Addresses (IPv4 / IPv6)IPv4, IPv6, ICMP, IPsec, ARP, OSPF, BGP
Layer 2Data LinkData LinkFramePhysical MAC Addresses (48-bit hex)Ethernet (802.3), Wi-Fi (802.11), PPP, VLAN (802.1Q)
Layer 1PhysicalPhysicalBitsVoltage levels, Light pulses, RF wavesCat5e/6/6a, Single/Multi-mode Fiber, RJ45, 1000BASE-T

2. Encapsulation & Decapsulation Mechanics

When a host sends data across a network, the transmission undergoes encapsulation as it traverses downward through the protocol stack from the application layer to the physical layer. Conversely, when the destination host receives the transmission, it undergoes decapsulation as it moves upward.

+-----------------------------------------------------------------------------------------+
|                           THE ENCAPSULATION PIPELINE                                    |
|                                                                                         |
|   [Application Layer]       +---------------------------------------------+             |
|                             | User Data (e.g., HTTP GET /index.html)      |             |
|                             +---------------------------------------------+             |
|                                                    |                                    |
|                                                    v Adds TCP/UDP Header                |
|   [Transport Layer]         +------------+--------------------------------+             |
|                             | TCP Header | Application Data Payload       |  (SEGMENT)  |
|                             +------------+--------------------------------+             |
|                                                    |                                    |
|                                                    v Adds IP Header                     |
|   [Network Layer]           +-----------+------------+--------------------+             |
|                             | IP Header | TCP Header | Application Data   |  (PACKET)   |
|                             +-----------+------------+--------------------+             |
|                                                    |                                    |
|                                                    v Adds Ethernet Header & Trailer     |
|   [Data Link Layer]   +----------+-----------+------------+------------------+--------+ |
|                       | Eth Header| IP Header | TCP Header | Application Data | Eth FCS| |
|                       +----------+-----------+------------+------------------+--------+ |
|                                                    |                     (FRAME)        |
|                                                    v Converts to bitstream              |
|   [Physical Layer]    0 1 1 0 1 0 0 1 0 1 1 0 1 1 1 0 0 0 1 0 1 1 0 1 0 0 1 1 (BITS)    |
+-----------------------------------------------------------------------------------------+

The Encapsulation Steps:

  1. Data Generation (Layer 5/7): The client application (e.g., a web browser) creates an application payload (e.g., an HTTP GET request).
  2. Transport Encapsulation (Layer 4): The OS transport stack appends a TCP header containing the source port (an ephemeral port such as 51234) and the destination port (80 or 443), along with sequence and acknowledgment numbers. This unit is a Segment.
  3. Network Encapsulation (Layer 3): The network stack appends an IP header containing the source IP address (192.168.1.50) and destination IP address (93.184.216.34), TTL (Time to Live), and protocol indicator (6 for TCP). This unit is a Packet.
  4. Data Link Encapsulation (Layer 2): The Network Interface Card (NIC) driver wraps the packet in an Ethernet frame, adding a header with the source MAC address and destination MAC address (the default gateway router's MAC if sending off-subnet), along with an EtherType field (0x0800 for IPv4). It also appends a Frame Check Sequence (FCS) trailer containing a 32-bit Cyclic Redundancy Check (CRC) hash for error detection. This unit is a Frame.
  5. Physical Transmission (Layer 1): The physical transceiver converts the digital frame into binary electrical pulses, light flashes, or RF signals for transmission across the medium.

Decapsulation & Integrity Verification:

Upon receipt, the destination device's NIC inspects the frame, re-calculates the CRC checksum, and compares it against the FCS trailer. If the checksum matches, the frame header and trailer are stripped (decapsulated), and the inner packet is passed up to Layer 3. The OS inspects the IP header, strips it, passes the segment to Layer 4, verifies port numbers, reassembles TCP streams, and delivers the clean application payload to the waiting application process.

Loading diagram...
Protocol Data Encapsulation and Decapsulation Flow

3. Physical Transmission Media: Copper, Fiber, and Coaxial

Physical Layer infrastructure forms the literal foundation of every computer network. Selecting the correct transmission media requires balancing bandwidth requirements, distance limitations, electromagnetic immunity, and installation costs.

+-----------------------------------------------------------------------------------------+
|                           TWISTED PAIR PINOUT COMPARISON                                |
|                                                                                         |
|   Pin   T568A Standard Wiring                     T568B Standard Wiring                 |
|   ---   ---------------------                     ---------------------                 |
|    1    White / Green                             White / Orange                        |
|    2    Green                                     Orange                                |
|    3    White / Orange                            White / Green                         |
|    4    Blue                                      Blue                                  |
|    5    White / Blue                              White / Blue                          |
|    6    Orange                                    Green                                 |
|    7    White / Brown                             White / Brown                         |
|    8    Brown                                     Brown                                 |
|                                                                                         |
|   [Note: T568A and T568B swap Pairs 2 (Orange) and 3 (Green); Pins 4,5,7,8 are identical]|
+-----------------------------------------------------------------------------------------+

Twisted Pair Copper Cabling

Twisted pair cabling contains four pairs of color-coded copper wires (8 conductors total). Wires are twisted together in pairs to exploit the principle of differential signaling and cancellation: external Electromagnetic Interference (EMI) affects both adjacent wires equally, allowing the receiving differential amplifier to cancel out the common-mode noise and eliminate internal crosstalk (signal bleed between adjacent wire pairs).

  • UTP (Unshielded Twisted Pair): Standard, flexible, inexpensive cabling used in most office LANs. Relies entirely on pair twisting for EMI rejection.
  • STP (Shielded Twisted Pair): Incorporates metallic foil shielding around individual pairs and/or an overall braided shielding jacket. Used in industrial settings, data centers, and environments with heavy machinery or RF interference.
  • Wiring Standards (TIA/EIA-568):
    • T568A vs. T568B: Two standardized termination pinout specifications for RJ45 (8P8C) modular connectors. T568B is the dominant standard in commercial enterprise installations in North America, while T568A is common in residential and government installations.
    • Straight-Through Cable: Terminated with the same standard on both ends (T568B to T568B). Used to connect disparate devices (e.g., Host to Switch, Switch to Router).
    • Crossover Cable: Terminated with T568A on one end and T568B on the other. Historically used to connect similar devices (e.g., Switch to Switch, Host to Host). Modern devices support Auto-MDIX (Automatic Medium-Dependent Interface Crossover), automatically detecting and configuring the transmit/receive pins regardless of cable pinout.

Ethernet Cable Categories Matrix

CategoryMaximum Bandwidth / Data RateMaximum FrequencyMaximum Standard DistanceTypical Use Case & Shielding
Cat5100 Mbps (Fast Ethernet)100 MHz100 meters (328 ft)Obsolete legacy installations; UTP
Cat5e1 Gbps (1000BASE-T)100 MHz100 meters (328 ft)Standard gigabit desktop drops; enhanced crosstalk specs; UTP
Cat610 Gbps (10GBASE-T) / 1 Gbps250 MHz55m at 10 Gbps; 100m at 1 GbpsModern commercial office cabling; central spline separator; UTP/STP
Cat6a10 Gbps (10GBASE-T)500 MHz100 meters (328 ft)Enterprise data centers, 10G backbones; augmented shielding; STP/UTP
Cat710 Gbps600 MHz100 meters (328 ft)Proprietary shielding (S/FTP); GG45/TERA connectors; data centers
Cat825 Gbps / 40 Gbps2000 MHz (2 GHz)30 meters (98 ft)Data center Top-of-Rack (ToR) switch-to-server interconnects; fully shielded

Fiber Optic Cabling

Fiber optic cables transmit pulses of light through flexible, ultra-pure silica glass or plastic cores. Because light does not conduct electricity, fiber optics are 100% immune to electromagnetic interference (EMI), radio frequency interference (RFI), and lightning surges, making them the standard for building-to-building campus backbones and telecommunications infrastructure.

  • Single-Mode Fiber (SMF):
    • Core Diameter: Extremely narrow (~9 microns / um).
    • Light Source: Solid-state Infrared Laser.
    • Optical Characteristics: Light travels along a single direct path without bouncing, eliminating modal dispersion (signal blurring caused by light rays taking different length paths).
    • Distance & Speed: Supports 10G/40G/100G+ over distances up to 40 to 100+ kilometers.
    • Use Case: Long-haul telecommunications, metropolitan area networks (MANs), WAN backbones.
  • Multi-Mode Fiber (MMF):
    • Core Diameter: Wider core (~50 or 62.5 microns / um).
    • Light Source: Inexpensive Light Emitting Diodes (LEDs) or Vertical-Cavity Surface-Emitting Lasers (VCSELs).
    • Optical Characteristics: Light rays bounce along the core walls across multiple optical modes, introducing modal dispersion that limits transmission distance.
    • Distance & Speed: Supports 1 Gbps to 100 Gbps up to 300 to 550 meters.
    • Use Case: Intra-building LAN backbones, data center server-to-switch interconnects.
  • Common Fiber Connectors:
    • LC (Lucent Connector / Local Connector): Compact, high-density form factor with a push-pull latching tab (dominant in SFP/SFP+ transceiver modules).
    • SC (Subscriber Connector / Standard Connector): Square push-pull snap connector.
    • ST (Straight Tip): Round bayonet-style twist-and-lock connector (common in older legacy MMF installations).

Coaxial Cabling

Coaxial cable features a central solid copper conductor surrounded by a dielectric plastic insulator, a woven metallic shielding braid, and an outer PVC jacket. Used primarily for broadband cable internet (DOCSIS standard) and closed-circuit television (CCTV). Terminated using F-type screw-on connectors (cable modems) or BNC bayonet connectors.


4. Network Hardware & Domain Segmentation

Connecting devices into scalable networks requires active hardware devices operating at different OSI layers. Understanding the boundaries of collision domains and broadcast domains is essential for diagnosing network congestion and designing robust topologies.

+-----------------------------------------------------------------------------------------+
|                       COLLISION VS. BROADCAST DOMAIN SEGMENTATION                       |
|                                                                                         |
|   [HUB (Layer 1)]       All connected ports share 1 COLLISION domain                    |
|                         and 1 BROADCAST domain. Collisions occur frequently.             |
|                                                                                         |
|   [SWITCH (Layer 2)]    Each switch port is an independent COLLISION domain (Microseg). |
|                         All ports still share a single BROADCAST domain (unless VLANs). |
|                                                                                         |
|   [ROUTER (Layer 3)]    Each router interface breaks BOTH Collision Domains             |
|                         AND Broadcast Domains. Does not forward broadcasts.             |
+-----------------------------------------------------------------------------------------+

Hubs (Layer 1 - Physical)

  • Operational Model: Multi-port physical repeaters. When a bit enters one port, the hub regenerates the electrical signal and broadcasts it out of every other port, regardless of destination.
  • Collision Domain Impact: All connected devices share a single collision domain.
  • Performance & Security: Hosts must operate in half-duplex mode using CSMA/CD (Carrier Sense Multiple Access with Collision Detection). If two hosts transmit simultaneously, an electrical collision occurs, requiring both hosts to back off for a random interval before retransmitting. Hubs are obsolete and pose severe security risks, as any connected packet sniffer can capture all network traffic.

Switches (Layer 2 - Data Link)

  • Operational Model: Intelligent multiport bridges that forward traffic based on hardware MAC addresses. A switch maintains a MAC Address Table (stored in high-speed Content Addressable Memory / CAM).
  • Switch Processing Cycle (Learning, Flooding, Forwarding, Filtering):
    1. Learning: When a frame enters a port, the switch inspects the source MAC address and associates that MAC with the ingress physical port in its CAM table.
    2. Flooding (Unknown Unicast): If the destination MAC address is not yet in the CAM table, the switch floods the frame out all ports except the receiving port.
    3. Forwarding: Once the destination MAC is learned, the switch forwards frames destined for that MAC only to the specific corresponding port.
    4. Filtering: Frames destined for a device on the same port or invalid frames are dropped/filtered.
  • Microsegmentation: Every single switch port constitutes its own isolated collision domain, enabling full-duplex communication (simultaneous transmitting and receiving at wire speed with zero collisions).
  • VLANs (Virtual LANs - IEEE 802.1Q): Switches can logically partition ports into distinct virtual networks, breaking a large physical switch into multiple isolated broadcast domains.

Routers (Layer 3 - Network)

  • Operational Model: Multi-interface internetworking devices that inspect Layer 3 IP headers to route packets across separate IP networks and subnets. Routers evaluate destination IP addresses against a Routing Table to choose the optimal next-hop interface.
  • Domain Boundary: Routers break broadcast domains. By default, routers do not forward Layer 2 broadcast frames (destination MAC FF:FF:FF:FF:FF:FF) or Layer 3 IP broadcasts (255.255.255.255), isolating broadcast storms to their local subnet.

Comprehensive Network Device Comparison Matrix

DevicePrimary OSI LayerHardware Addressing UsedCollision Domain BoundaryBroadcast Domain BoundaryDuplex SupportTypical Function
HubLayer 1 (Physical)None (Blind repeating)1 shared domain across all ports1 shared domainHalf-duplex onlyObsolete legacy signal repeating
BridgeLayer 2 (Data Link)MAC Addresses (Software)Breaks collision domains per port (few ports)1 shared domainFull / HalfLegacy two-segment LAN interconnection
SwitchLayer 2 (Data Link)MAC Addresses (CAM Table)Independent collision domain per port1 shared domain (per VLAN)Full-duplex dedicatedHigh-density LAN switching & microsegmentation
RouterLayer 3 (Network)IP Addresses (Routing Table)Independent collision domain per interfaceBreaks broadcast domains per interfaceFull-duplexInter-network routing & WAN boundary gateway
Multilayer SwitchLayers 2 & 3MAC + IP AddressesIndependent collision domain per portBreaks broadcast domains per VLAN interfaceFull-duplexHigh-performance wire-speed inter-VLAN routing
Test Your Knowledge

An IT technician is analyzing a captured network transmission and observes the following headers: an Ethernet frame header with source and destination MAC addresses, encapsulating an IP header with source and destination IP addresses, which in turn encapsulates a TCP header with port numbers. What is the correct term for the Protocol Data Unit (PDU) at the Network layer before Data Link encapsulation occurs?

A
B
C
D
Test Your Knowledge

A network engineer needs to install horizontal copper cabling in a new office building to support 10 Gbps Ethernet throughput over a distance of 85 meters. Which minimum cable category must be specified to meet this requirement?

A
B
C
D
Test Your Knowledge

An unmanaged 24-port Layer 2 Ethernet switch has 20 active computers connected with no VLANs configured. How many collision domains and how many broadcast domains exist on this switch?

A
B
C
D
Test Your Knowledge

Which of the following optical fiber characteristics is specific to Single-Mode Fiber (SMF) when compared to Multi-Mode Fiber (MMF)?

A
B
C
D