Network Hardware: Physical Components of Networked Systems
Network Devices
Read aloud is unavailable in this browser.
The Active Components That Make Networks Function
Cables carry signals. Network interface cards translate between computers and cables. But it's network devices—the active, intelligent equipment deployed throughout a network—that actually make large-scale, efficient, secure networking possible.
Without switches, every device on a network would hear every other device's traffic. Without routers, the Internet would be an unnavigable chaos of isolated networks. Without firewalls, every system would be exposed to attack. Each category of network device solves specific problems that arise when connecting multiple computers, and understanding these devices is essential for anyone designing, operating, or troubleshooting networks.
By the end of this page, you will understand the purpose and operation of major network device categories: hubs, switches, routers, wireless access points, firewalls, load balancers, and more. You'll learn how each operates at specific OSI layers, the differences between layer 2 and layer 3 forwarding, and how modern networks combine these functions into sophisticated, software-defined platforms.
Network Device Hierarchy and OSI Layer Operation
Network devices operate at different layers of the OSI model, and understanding this relationship is fundamental to understanding their capabilities and limitations.
The Layer Principle:
A device operating at a given layer can read and act on information from that layer and all layers below it, but typically treats higher-layer data as opaque payload.
| OSI Layer | Device Category | Key Function |
|---|---|---|
| Layer 1 (Physical) | Repeater, Hub | Signal regeneration, physical connectivity |
| Layer 2 (Data Link) | Switch, Bridge | Frame forwarding based on MAC addresses |
| Layer 3 (Network) | Router, L3 Switch | Packet routing based on IP addresses |
| Layer 4 (Transport) | Firewall, Load Balancer | Port-based filtering, connection management |
| Layer 7 (Application) | Application Firewall, Proxy | Content inspection, application-aware routing |
Key Distinctions:
Layer 2 Devices:
- Forward frames based on MAC addresses
- Learn addresses from traffic (no configuration required for basic operation)
- Cannot cross network boundaries (single broadcast domain)
- Examples: Switches, Bridges, Wireless Access Points (at L2)
Layer 3 Devices:
- Forward packets based on IP addresses
- Require configuration of network/subnet information
- Connect different networks, enabling inter-network communication
- Examples: Routers, Layer 3 Switches, Gateways
Multi-Layer Devices:
Modern network devices often operate at multiple layers simultaneously:
- Layer 3 Switch: Performs both L2 switching and L3 routing in hardware
- Next-Generation Firewall (NGFW): L3/L4 filtering plus L7 application inspection
- Software Load Balancer: L4 connection distribution plus L7 content routing
Historically, routers routed and switches switched. Today, high-end switches can route at wire speed, routers include powerful firewall features, and software-defined solutions blur hardware boundaries entirely. Understanding the functions matters more than rigid device classifications.
Hubs and Repeaters: Layer 1 Devices (Historical Context)
Repeaters and hubs are the simplest network devices, operating entirely at the physical layer. While largely obsolete in modern wired networks, understanding them provides important context for networking concepts.
Repeater:
A repeater receives a signal on one port, regenerates it (cleaning up any degradation), and retransmits it on another port.
- Purpose: Extend distance beyond a single cable segment's limit
- Operation: Pure signal regeneration; has no awareness of data content
- Limitation: Does not filter or manage traffic; amplifies noise along with signals
Hub (Multi-Port Repeater):
A hub is essentially a multi-port repeater. A signal received on any port is regenerated and sent out all other ports.
- Simple connectivity: Connect multiple devices easily
- Shared bandwidth: All ports share the same collision domain
- No intelligence: Cannot direct traffic; all data goes everywhere
- •Single collision domain — Only one device can transmit at a time
- •Shared bandwidth — 100 Mbps hub means 100 Mbps total, not per port
- •Security exposure — All traffic visible to all connected devices
- •Collision overhead — Higher traffic = more collisions = worse performance
- •Half-duplex only — Cannot send and receive simultaneously
- •Per-port collision domains — Multiple simultaneous transmissions
- •Dedicated bandwidth — Each port gets full bandwidth
- •Traffic isolation — Unicast frames go only to destination port
- •No collision overhead — Modern switches operate collision-free
- •Full-duplex — Simultaneous bidirectional communication
Hubs are nearly extinct in production networks, but the concepts they illustrate—collision domains, shared media, broadcast behavior—remain fundamental. These concepts apply to wireless networks (which are inherently shared media) and help explain why switches were such a revolutionary improvement.
Switches: Layer 2 Forwarding and MAC Learning
Switches are the workhorses of modern local area networks. They operate at Layer 2 (Data Link), making forwarding decisions based on MAC addresses. Unlike hubs, switches create separate collision domains per port while maintaining a shared broadcast domain.
Fundamental Switch Operation:
1. MAC Address Learning:
When a frame arrives on a port, the switch examines the source MAC address and records it in the MAC address table (also called CAM table—Content Addressable Memory):
MAC Address Table Entry:
┌─────────────────────┬──────────┬────────────────┐
│ MAC Address │ Port │ Timer │
├─────────────────────┼──────────┼────────────────┤
│ 00:1A:2B:3C:4D:5E │ Gi0/1 │ 300 seconds │
│ AA:BB:CC:DD:EE:FF │ Gi0/24 │ 287 seconds │
└─────────────────────┴──────────┴────────────────┘
2. Frame Forwarding:
When forwarding a frame, the switch looks up the destination MAC address:
- Known unicast: Frame is sent only to the port where that MAC was learned
- Unknown unicast: Frame is flooded to all ports except the source port
- Broadcast (FF:FF:FF:FF:FF:FF): Frame is flooded to all ports
- Multicast: Flooded unless IGMP snooping is configured
3. Filtering:
If the source and destination are on the same port, the frame is filtered (not forwarded), as the devices can communicate directly.
Switch Features and Capabilities:
VLANs (Virtual LANs):
VLANs partition a physical switch into multiple logical switches:
- Port-based VLANs: Ports assigned to different VLANs
- Tagged VLANs (802.1Q): Frames carry VLAN ID tags for multi-switch environments
- Benefits: Traffic isolation, security segmentation, broadcast domain control
Spanning Tree Protocol (STP):
Prevents loops in networks with redundant paths:
- Elects a root bridge (lowest bridge ID)
- Calculates shortest paths to root
- Blocks redundant paths to prevent loops
- Reconverges if topology changes
- Variants: STP, RSTP (Rapid), MSTP (Multiple)
Link Aggregation (LAG/LACP):
Combines multiple physical links into one logical link:
- Increased bandwidth (additive)
- Redundancy if one link fails
- IEEE 802.3ad / 802.1AX standard
Port Mirroring (SPAN):
Copies traffic from one or more ports to a monitoring port:
- Used for network analysis and troubleshooting
- IDS/IPS integration
- Compliance monitoring
| Type | Management | Features | Use Case |
|---|---|---|---|
| Unmanaged | None | Basic L2 switching only | Home, small office, workgroup |
| Smart/Web-managed | Basic web UI | VLANs, QoS, limited features | SMB, non-critical deployments |
| Managed L2 | Full CLI/GUI/SNMP | VLANs, STP, LAG, security | Enterprise access/distribution |
| Managed L3 | Full management | L2 + IP routing, ACLs | Enterprise core, data center |
| Data Center | SDN integration | High density, low latency, automation | Cloud, hyperscale, HPC |
Switches have finite MAC table capacity (typically 8K-128K entries). An attacker flooding a switch with fake MAC addresses can exhaust this table, causing the switch to fall back to hub-like behavior (flooding everything). Port security features limit MAC addresses per port to mitigate this attack.
Routers: Layer 3 Packet Forwarding and Routing
Routers are the fundamental building blocks of internetworking. They operate at Layer 3 (Network), making forwarding decisions based on IP addresses. Every packet that crosses network boundaries—including every packet on the Internet—is routed.
Router vs. Switch: Fundamental Differences:
| Aspect | Switch (L2) | Router (L3) |
|---|---|---|
| Forwarding basis | MAC addresses | IP addresses |
| Broadcast handling | Forwards broadcasts | Blocks broadcasts (by default) |
| Domain | Single broadcast domain | Separates broadcast domains |
| Address learning | Automatic from traffic | Requires configuration |
| Table type | MAC/CAM table | Routing/FIB table |
| Typical latency | < 5 µs | 10-100 µs |
Routing Table:
The routing table tells the router how to reach different networks:
Routing Table Example:
┌─────────────────────┬───────────────┬─────────────┬────────────┐
│ Destination Network │ Next Hop │ Interface │ Metric │
├─────────────────────┼───────────────┼─────────────┼────────────┤
│ 10.0.0.0/8 │ Directly Conn │ GigE0/0 │ 0 │
│ 192.168.1.0/24 │ 10.0.0.1 │ GigE0/0 │ 1 │
│ 172.16.0.0/16 │ 10.0.0.2 │ GigE0/0 │ 2 │
│ 0.0.0.0/0 │ 203.0.113.1 │ GigE0/1 │ 1 (default)│
└─────────────────────┴───────────────┴─────────────┴────────────┘
Packet Forwarding Process:
- Receive packet on incoming interface
- Decrement TTL (Time to Live); drop if TTL = 0
- Lookup destination IP in routing table
- Find longest prefix match (most specific route)
- Determine next hop and egress interface
- Perform ARP lookup for next hop's MAC address (if not cached)
- Rewrite Ethernet header with new source/destination MACs
- Recalculate frame CRC
- Transmit on egress interface
Routing Protocols:
Routers populate their routing tables via:
Static Routes:
- Manually configured by administrators
- Simple and predictable
- Doesn't adapt to topology changes
- Used for default routes, stub networks
Dynamic Routing Protocols:
| Protocol | Type | Algorithm | Typical Use |
|---|---|---|---|
| RIP | Distance Vector | Bellman-Ford | Legacy, small networks |
| OSPF | Link State | Dijkstra SPF | Enterprise, data center |
| IS-IS | Link State | Dijkstra SPF | Large ISPs, data centers |
| EIGRP | Hybrid | DUAL | Cisco environments |
| BGP | Path Vector | Path selection | Internet backbone, multi-homed |
- •Access Control Lists (ACLs) — Filter packets based on source/destination IP, ports, protocols
- •Network Address Translation (NAT) — Translate between private and public IP addresses
- •Quality of Service (QoS) — Prioritize certain traffic types (voice, video, etc.)
- •VPN Termination — Establish encrypted tunnels (IPsec, GRE, etc.)
- •Policy-Based Routing — Route based on criteria beyond destination address
- •DHCP Relay — Forward DHCP requests across network boundaries
- •First-Hop Redundancy — VRRP, HSRP for gateway failover
Layer 3 switches route in dedicated hardware (ASICs) at wire speed, while traditional routers often use software-based forwarding. L3 switches excel at high-speed, simple routing within data centers. Traditional routers provide richer features for WAN connectivity, VPNs, and complex routing policies.
Wireless Access Points: Bridging Wired and Wireless
Wireless Access Points (APs) enable wireless devices to connect to wired networks. They operate at Layer 2, bridging the 802.11 (Wi-Fi) and 802.3 (Ethernet) domains. As wireless networking becomes ubiquitous, understanding AP architecture and deployment is essential.
Basic AP Operation:
- Beacon transmission: AP periodically broadcasts its presence (SSID, capabilities)
- Association: Wireless client discovers AP and associates with it
- Authentication: Security protocols verify client identity (WPA2/WPA3)
- Frame bridging: AP converts 802.11 frames to 802.3 and vice versa
- Channel management: AP coordinates simultaneous transmissions
AP Deployment Models:
| Model | Control Plane | Data Plane | Best For |
|---|---|---|---|
| Standalone (Fat AP) | Local on AP | Local on AP | Home, SOHO, simple deployments |
| Controller-based (Thin AP) | Centralized controller | Through controller or local | Enterprise, large-scale |
| Cloud-managed | Cloud controller | Local on AP | Distributed sites, managed services |
| Mesh | Distributed/controller | Wireless backhaul | Large areas, difficult wiring |
Wi-Fi Standards Evolution:
| Generation | Standard | Max Rate | Frequency | Key Features |
|---|---|---|---|---|
| Wi-Fi 4 | 802.11n | 600 Mbps | 2.4/5 GHz | MIMO, channel bonding |
| Wi-Fi 5 | 802.11ac | 3.5 Gbps | 5 GHz only | MU-MIMO (down), 80/160 MHz |
| Wi-Fi 6 | 802.11ax | 9.6 Gbps | 2.4/5 GHz | OFDMA, TWT, BSS coloring |
| Wi-Fi 6E | 802.11ax | 9.6 Gbps | 2.4/5/6 GHz | 6 GHz band (1200 MHz more) |
| Wi-Fi 7 | 802.11be | 46 Gbps | 2.4/5/6 GHz | 320 MHz, 4K QAM, MLO |
Enterprise AP Considerations:
Radio Resource Management (RRM):
- Automatic channel selection to minimize interference
- Transmit power adjustment for coverage optimization
- Client load balancing across APs
Security Features:
- WPA3-Enterprise with 802.1X authentication
- Wireless intrusion detection/prevention (WIDS/WIPS)
- Rogue AP detection
- Client isolation
Quality of Service:
- WMM (Wi-Fi Multimedia) prioritization
- Airtime fairness
- Bandwidth limits per client/SSID
Unlike wired networks where cables determine connectivity, wireless coverage depends on AP placement, building materials, and RF environment. Professional site surveys using spectrum analyzers and predictive modeling tools are essential for enterprise deployments. Poor placement leads to dead zones, interference, and poor user experience.
Firewalls and Security Appliances
Firewalls are security devices that control traffic flow between networks based on configured rules. They form the primary perimeter defense for most organizations and have evolved from simple packet filters to sophisticated, multi-function security platforms.
Firewall Evolution:
Generation 1: Packet Filtering (Stateless)
- Inspects individual packets in isolation
- Rules based on source/dest IP, ports, protocol
- No connection awareness
- Fast but easily bypassed
Generation 2: Stateful Inspection
- Tracks connection state (TCP handshake, established, etc.)
- Allows return traffic for established connections
- Blocks packets that don't match expected state
- The standard for decades
Generation 3: Application-Aware (NGFW)
- Deep packet inspection (DPI)
- Identifies applications regardless of port
- User identity integration
- Integrated IPS, malware detection, URL filtering
| Capability | Packet Filter | Stateful | NGFW |
|---|---|---|---|
| IP/Port filtering | ✓ | ✓ | ✓ |
| Connection tracking | ✗ | ✓ | ✓ |
| Application identification | ✗ | ✗ | ✓ |
| User identity awareness | ✗ | ✗ | ✓ |
| Integrated IPS | ✗ | ✗ | ✓ |
| TLS decryption | ✗ | ✗ | ✓ |
| Malware inspection | ✗ | ✗ | ✓ |
| URL/Category filtering | ✗ | Limited | ✓ |
Other Security Devices:
Intrusion Detection/Prevention Systems (IDS/IPS):
- IDS: Monitors and alerts on suspicious activity
- IPS: Can actively block detected threats
- Uses signature-based and anomaly-based detection
- Typically deployed inline (IPS) or via port mirror (IDS)
Web Application Firewall (WAF):
- Protects web applications from layer 7 attacks
- Defends against SQL injection, XSS, CSRF, etc.
- Often deployed in front of web servers
Network Access Control (NAC):
- Verifies device compliance before network access
- Checks posture: antivirus, patches, configuration
- Quarantines non-compliant devices
DDoS Mitigation Appliances:
- Detects and absorbs volumetric attacks
- Behavioral analysis to identify attack traffic
- Rate limiting and traffic scrubbing
- Often cloud-based for scale
A firewall is one component of defense-in-depth, not a complete security solution. Encrypted traffic, authorized user actions, zero-day exploits, and insider threats can all bypass firewall controls. Modern security requires endpoint protection, monitoring, identity management, and security awareness alongside network controls.
Load Balancers and Application Delivery Controllers
Load balancers distribute incoming traffic across multiple backend servers, providing scalability, redundancy, and optimal resource utilization. Modern Application Delivery Controllers (ADCs) extend this with advanced traffic management, security, and optimization features.
Why Load Balancing Matters:
- Scalability: Handle more traffic than a single server can manage
- Availability: If one server fails, others continue serving requests
- Performance: Distribute load evenly; prevent any single server from becoming overwhelmed
- Flexibility: Add/remove servers without user impact
Load Balancing Layers:
- •Distributes based on IP and port
- •No inspection of application content
- •Very fast, low overhead
- •Works with any TCP/UDP application
- •Persistence via source IP
- •Example: DNS, TCP passthrough
- •Inspects application content (HTTP headers, URLs)
- •Content-based routing decisions
- •More CPU intensive
- •Advanced features (SSL termination, compression)
- •Session persistence via cookies
- •Example: HTTP/HTTPS load balancing
Load Balancing Algorithms:
| Algorithm | Description | Best For |
|---|---|---|
| Round Robin | Rotate through servers sequentially | Equal-capacity servers, stateless apps |
| Weighted Round Robin | Round robin with capacity weights | Mixed-capacity servers |
| Least Connections | Send to server with fewest active connections | Long-lived connections, varied load |
| Weighted Least Connections | Least connections with capacity weighting | Mixed capacity + varied connections |
| IP Hash | Hash client IP to select server | Session persistence without cookies |
| URL Hash | Hash request URL to select server | Caching optimization |
| Least Response Time | Send to fastest-responding server | Latency-sensitive applications |
ADC Advanced Features:
- SSL/TLS Termination: Offload encryption from backend servers
- Content Caching: Store and serve static content
- Compression: Reduce bandwidth for text content
- Connection Multiplexing: Many client connections → fewer backend connections
- Health Checking: Monitor backend servers; remove unhealthy servers from pool
- Global Server Load Balancing (GSLB): Distribute across geographically dispersed data centers
- Rate Limiting: Protect backends from traffic spikes
- WAF Integration: Combine load balancing with application security
Cloud platforms provide managed load balancing services (AWS ALB/NLB, Azure Load Balancer, GCP Load Balancing) that integrate with auto-scaling and other cloud services. These are often preferred over hardware appliances in cloud environments, though the fundamental concepts remain identical.
KEEP LEARNING
There’s more to understand.
You’re reading the introduction. Create a free account to read the complete lesson, save notes and keep track of your progress.