A Brief History of Network Hardware: From Cables to Modern Switches

Network fundamentals

Before networking, personal computers operated independently without network cards, cables, or protocol stacks, using physical media such as disks or CDs for data transfer.

When Ethernet cables appeared, the smallest network unit consisted of a cable (physical medium for bit‑stream transmission), a network interface card (converts stored data to electrical signals on the cable), and a protocol stack (provides addressing, flow control, and data analysis).

Physical‑layer extensions

If the distance between endpoints exceeds the cable’s physical limit, data loss occurs.

Repeater : a physical‑layer device that amplifies and relays signals to extend reach.

Hub : a multi‑port repeater that broadcasts incoming frames to all other ports; limited to two ports for direct host‑to‑host communication.

Data‑link‑layer devices

Bridge acts as a “smart” repeater: it extends physical connectivity and partitions MAC addresses to isolate collision domains, reducing unnecessary traffic.

Switch builds on bridge functionality by maintaining a MAC address table (a map) that forwards frames only to the appropriate destination. Compared with bridges, switches provide:

Denser port density—each host resides in its own collision domain, greatly improving bandwidth utilization.

ASIC‑based hardware for high‑speed forwarding.

VLAN isolation, which separates both collision and broadcast domains.

Switches operate within LAN environments and do not provide WAN routing.

Network‑layer routing

The first router was created by Leonard Bosack and Sandy Lerner at Stanford for the SUNet campus network. It uses IP addressing and routing tables to forward packets, connecting different LANs, isolating broadcast domains, and enabling WAN communication.

Logical addressing mechanisms such as the IP protocol are essential for inter‑LAN connectivity; hosts with proper IP addresses and subnet planning can communicate across disparate LANs.

Wireless access points (AP)

Wireless APs combine switch/router functionality with Wi‑Fi radio. Two deployment models exist:

Fat AP : contains an independent operating system and can be configured locally (similar to home‑grade APs).

Thin AP : provides only radio transmission; all configuration is centralized on a controller (AC). Fat APs suit small‑scale networks (home, small office); thin APs are used for large‑scale deployments (wireless cities, campus networks).

Network security devices

Firewall limits network access at the edge to protect against external attacks. Types include:

Packet‑filtering firewalls.

Application‑proxy firewalls.

Stateful inspection firewalls.

Firewalls can be software or hardware appliances. Early firewalls added access‑control functions on top of routers; modern firewalls may be placed before or after routers, separating routing (address translation, routing policies) from security (traffic isolation). Extensions such as Web firewalls, security gateways, and IDS/IPS are built on the firewall foundation.

Traffic‑control devices

To mitigate network congestion, traffic‑control devices are categorized as:

Behavior‑management : fine‑grained traffic classification and control.

Load balancer / Application Delivery Controller (ADC) : distributes traffic based on characteristics (application, address, etc.) across links and servers.

Link‑optimization : improves utilization of low‑speed WAN links by maximizing throughput.

Typical network topologies and required devices

Home SOHO network : wireless router provides Wi‑Fi access and external connectivity.

Small‑business network : two‑layer architecture with a router, switch, and server.

Campus/enterprise network : three‑layer (access‑aggregation‑core) design with multiple switches, routers, and firewalls; zones include user, service, external, management, and Internet.

Data‑center network : second‑layer design with tenant zones (virtualized compute and network), Internet zones (BGP routers, IPS/anti‑DDoS), and security management zones (firewalls, audit, IDS). Load balancers reside in the server tier to distribute traffic among fixed servers.

These layers illustrate the problem‑solution chain: a networking limitation (cable length, port scarcity, security, congestion) leads to the introduction of a specific hardware device that addresses the issue, often with trade‑offs between functionality, scalability, and deployment complexity.

Code example

来源：弱电文档