Local Area Networks (LAN)

A Local Area Network is more than just 'computers connected together'—it's a carefully engineered system of interconnected components, each serving specific functions. From the network interface card inside your laptop to the fiber optic backbone connecting buildings across a campus, every component plays a critical role in delivering reliable, high-speed connectivity.

Understanding LAN components isn't just academic knowledge—it's practical expertise for:

• Network Design: Selecting appropriate components for performance, reliability, and budget
• Troubleshooting: Isolating failures to specific components for rapid resolution
• Capacity Planning: Understanding component capabilities and limitations for growth
• Security Implementation: Knowing where security controls can be applied in the infrastructure

This page provides comprehensive coverage of physical and active components, from passive cabling to intelligent switching systems.

The Network Interface Card (NIC) is the point where a device connects to the network. Every computer, server, printer, IoT device, or any network-capable equipment contains at least one NIC—whether as a discrete card, integrated into the motherboard, or as a USB adapter.

Physical Form Factors:

NIC Functions:

A NIC performs several critical functions:

1. Physical Layer Operations:

• Convert digital data to electrical/optical signals for transmission
• Convert received signals back to digital data
• Clock recovery and signal synchronization
• Line coding (encoding bytes into transmission symbols)

2. Data Link Layer Operations:

• Frame construction with proper headers, payload, and FCS
• MAC address handling (burned-in address, configurable overrides)
• Error detection via FCS calculation and verification
• Flow control via pause frames (802.3x)

3. Advanced Features (Modern NICs):

• Checksum offload: Calculate/verify IP, TCP, UDP checksums in hardware
• Segmentation offload (TSO/LSO): Hardware handles large segment division
• Receive Side Scaling (RSS): Distribute incoming traffic across CPU cores
• VLAN tag insertion/removal: Hardware VLAN processing
• SR-IOV: Virtual NIC instances for VM direct access

NIC Performance Considerations:

Throughput vs. Packets Per Second: NIC performance has two dimensions:

• Throughput: Maximum data transfer rate (e.g., 10 Gbps)
• PPS (Packets Per Second): Maximum rate of packet processing

A NIC might achieve full 10 Gbps with large packets but struggle with millions of small packets per second. High-frequency trading and network functions require high PPS capability.

Interrupt Coalescing: Processing every packet as a separate CPU interrupt creates overhead. Modern NICs batch multiple packets into single interrupts, trading a small latency increase for significant CPU savings.

DMA and Ring Buffers: NICs use Direct Memory Access (DMA) to transfer data directly to system RAM without CPU involvement. Ring buffers (circular queues of packet descriptors) coordinate between the NIC and operating system.

Multi-Queue NICs: High-performance NICs have multiple transmit and receive queues, enabling parallel processing across CPU cores. This is essential for achieving multi-gigabit throughput in software-based packet processing.

Twisted-pair copper cabling is the dominant transmission medium for horizontal LAN cabling—the cables running from wiring closets to end-user devices. Understanding cabling categories, specifications, and installation practices is essential for LAN infrastructure.

Twisted Pair Basics:

Twisted-pair cables contain four pairs of copper wires, each pair twisted together. The twisting:

• Reduces electromagnetic interference (EMI) from external sources
• Reduces crosstalk between pairs within the same cable
• Creates a balanced transmission line that rejects common-mode noise

Cabling Categories:

UTP vs. STP:

• UTP (Unshielded Twisted Pair): No additional shielding; relies on twisting for noise rejection. Lower cost, easier to install, adequate for most environments.

• STP (Shielded Twisted Pair): Each pair and/or the entire cable has metallic shielding. Better noise immunity but higher cost, more difficult installation, and requires proper grounding.

F/UTP, S/FTP, and Other Variants: Modern shielding notation uses 'X/YTP' format:

• First letter: Overall cable shield (U = unshielded, F = foil, S = braid)
• Second letter: Individual pair shield (U = unshielded, F = foil)
• Example: F/UTP = foil around all pairs, no individual pair shields

The RJ-45 Connector:

The 8P8C modular connector (commonly called RJ-45) terminates twisted-pair Ethernet cables:

• 8 positions, 8 contacts
• Two wiring standards: T568A and T568B (differ in which pairs connect to which pins)
• Cables should use the same standard at both ends (straight-through)
• Crossover cables (A at one end, B at other) were historically needed for switch-switch connections, but Auto-MDI/MDIX has eliminated this requirement

Power Over Ethernet (PoE):

PoE delivers DC power over Ethernet cables to devices like IP phones, wireless access points, and security cameras:

PoE uses either spare pairs (Mode A) or the same pairs as data (Mode B, using phantom power), depending on the Ethernet speed and PoE generation.

Fiber optic cables transmit data as light pulses through glass or plastic strands. In LANs, fiber is essential for backbone connections, building-to-building links, and high-speed server connectivity. Fiber's immunity to electromagnetic interference and ability to span long distances make it indispensable.

Fiber Types:

Fiber Connectors:

LC connectors dominate modern deployments due to their small size (half the footprint of SC), enabling higher port density.

Fiber Termination:

Fiber can be terminated with:

• Pre-terminated assemblies: Factory-made cables with connectors installed; highest quality, requires planning
• Field termination: Connectors attached on-site using epoxy/polish or mechanical splice connectors
• Fusion splicing: Permanently joining fibers using heat; lowest loss but requires expensive equipment

Transceiver Modules:

Fiber connections typically use modular transceivers:

Structured cabling is a standardized approach to building telecommunications infrastructure. Rather than running random cables point-to-point, structured cabling creates a hierarchical, organized system that supports current needs and future growth.

TIA/EIA-568 Standards:

The Telecommunications Industry Association (TIA) defines structured cabling standards:

• TIA-568: Commercial building cabling standard
• TIA-569: Pathways and spaces (conduits, closets, equipment rooms)
• TIA-606: Administration (labeling and documentation)
• TIA-607: Grounding and bonding

Cabling Subsystems:

Structured cabling divides infrastructure into subsystems:

Patch Panels:

Patch panels provide organized termination points for horizontal cabling:

• Front: RJ-45 jacks for patch cables connecting to switches
• Rear: Permanent terminations (punch-down) for horizontal cables

This design isolates permanent infrastructure from frequently-changed connections. Horizontal cables are never directly connected to switches—changes happen at the patch panel using patch cables.

Channel vs. Permanent Link:

• Permanent Link: The fixed infrastructure (outlet → horizontal cable → patch panel termination)

  • Maximum: 90 meters
  • Tested during installation certification

• Channel: Complete end-to-end path including patch cables

  • Maximum: 100 meters total (90m permanent link + 5m equipment patch + 5m work area patch)
  • What users actually experience

Cable Management:

Proper cable management ensures reliability and serviceability:

• Horizontal cable managers between patch panel rows
• Vertical cable managers between racks
• Service loops for slack management
• Proper bend radius maintenance (typically 4× cable diameter minimum)
• Color coding for circuit identification
• Labels on both ends of every cable

Switches are the central active component of any LAN. They connect all devices, forward frames based on MAC addresses, and provide the intelligent fabric that makes modern Ethernet possible.

Switch Categories:

Key Switch Specifications:

Switching Capacity (Backplane Bandwidth): Total internal bandwidth available for frame forwarding. A 48-port Gigabit switch with full-duplex requires 96 Gbps minimum (48 × 1 Gbps × 2). Non-blocking switches can forward at wire speed on all ports simultaneously.

Forwarding Rate (Throughput): Measured in packets per second (pps) or million packets per second (Mpps). Calculate required rate from port count and minimum frame sizes:

• 48 Gigabit ports at 64-byte frames = 48 × 1,488,095 pps = 71.4 Mpps required

Port Types:

• Access ports: Connect end devices (1 GbE typical)
• Uplink ports: Connect to distribution switches (10/25 GbE)
• SFP/SFP+ slots: Modular ports for copper or fiber transceivers
• Stacking ports: Proprietary high-speed inter-switch connections

MAC Address Table Size: Number of MAC addresses the switch can learn. Enterprise switches typically support 16,000-128,000+ MAC addresses. Insufficient capacity causes table overflow, forcing the switch to flood unknown destinations.

Wireless Access Points (APs) extend the LAN to mobile devices by bridging wireless (802.11) clients to the wired Ethernet infrastructure. Modern enterprise wireless deployments are complex systems requiring careful planning for coverage, capacity, interference management, and security.

Access Point Types:

Wi-Fi Standards:

Enterprise AP Features:

Beyond the active networking devices, LANs require supporting infrastructure for power, environmental control, physical security, and management. These often-overlooked components are essential for reliable operation.

Power Infrastructure:

Uninterruptible Power Supply (UPS):

• Provides battery backup during power outages
• Protects equipment from surges, sags, and electrical noise
• Sizing: Calculate total load (VA) and required runtime
• Online (double-conversion) UPS provides cleanest power for sensitive equipment

Power Distribution Unit (PDU):

• Distributes power within racks
• Basic PDUs: Simple outlets
• Metered PDUs: Monitor power consumption
• Managed PDUs: Remote outlet control, per-outlet monitoring
• Automatic Transfer Switch (ATS) PDUs: Switch between power sources

Redundant Power:

• Dual power supplies in critical equipment
• A+B power feeds from separate circuits/UPS units
• Generator backup for extended outages

Racks and Enclosures:

• Standard 19-inch racks: Industry standard width for equipment mounting
• Rack height: Measured in 'U' units (1U = 1.75 inches/44.45mm)
• Full-height rack: 42U (approximately 6 feet)
• Cable management: Vertical and horizontal organizers integral to design
• Security: Locking doors, side panels, environmental seals

Console and Out-of-Band Management:

• Console servers: Aggregate serial console ports for remote access
• Out-of-band network: Separate management network not dependent on production
• Remote KVM: IP-based keyboard/video/mouse for physical console access
• IPMI/iLO/iDRAC: Server management interfaces independent of OS

Documentation and Labeling:

• Every cable labeled at both ends
• Rack elevation diagrams showing equipment placement
• Network diagrams (physical and logical)
• IP address management (IPAM) database
• Port mapping documentation
• Change management records

We've comprehensively explored the components that comprise a Local Area Network—from the network interface cards in devices to the supporting power and environmental infrastructure that keeps everything running.

What's Next:

Now that we understand LAN components, we'll explore LAN applications—the services and use cases that LANs enable. The next page examines file sharing, printing, voice over IP, video conferencing, and other applications that depend on robust LAN infrastructure.