Network Hardware Evolution: From Cables to Data Center Architectures

The article begins by describing how personal computers operated in isolation before networking, relying on physical media like disks and CDs for data transfer. The invention of the network cable, network interface card (NIC), and protocol stack formed the smallest network unit: cables transmit bit streams as electrical signals, NICs convert stored data into those signals, and the protocol stack provides communication rules for data analysis, addressing, and flow control.

Extending Distance: Repeaters

When cable length exceeds physical transmission limits, data loss occurs. The repeater , a physical-layer device, relays and amplifies signals to enable long-distance transmission between devices.

Connecting Multiple Hosts: Hubs

Repeaters typically have only two ports, limiting direct communication among three or more hosts. The hub (multi-port repeater) solves this by receiving data on any interface and broadcasting it to all other interfaces, still operating at the physical layer.

Selective Forwarding: Bridges

Hubs flood traffic everywhere, causing collisions. The bridge acts as a "smart" repeater: it learns MAC addresses, partitions the network into separate collision domains, and forwards frames only to the relevant segment.

High-Speed Switching: Switches

Switches evolved from bridges with three key improvements: (1) denser ports, placing each host in its own collision domain and greatly improving bandwidth utilization; (2) dedicated ASIC hardware chips for high-speed forwarding; (3) VLAN support to isolate broadcast domains in addition to collision domains. Switches build a MAC address table (a map) to forward frames intelligently, but remain LAN devices unsuitable for wide-area communication.

Interconnecting Networks: Routers

The first router was invented by Stanford professors Leonard Bossack and Santi Lerner for the Stanford University Network (SUNet) and Cisco. Routers operate at the network layer using IP addressing and routing tables to forward packets. They connect different LANs, isolate broadcast domains, and enable WAN connectivity. Logical addressing (IP) and proper subnet planning allow hosts on disparate LAN types to communicate. The router's emergence drove the internet's explosive growth by enabling cross-media, cross-region network integration.

Wireless Connectivity: Access Points

Wireless APs function as wireless switches/routers. Two deployment models exist: Fat APs have independent OSes and configure hotspots locally (like home TP-Link devices); Thin APs only transmit wireless signals, with all configuration centralized on a backend Wireless Controller (AC). Small networks (home, small business) use fat APs; large deployments (wireless cities, campus networks) require thin AP (AC + AP) architectures.

Security Enforcement: Firewalls

Firewalls restrict network access, typically placed at the internet edge to block external attacks. They can be categorized by technology: packet filtering , application proxy , and stateful inspection ; and by form: software and hardware firewalls. Early firewalls added access control to routers, so routers often include firewall features like routing protocols, ACLs, and NAT. Firewalls and routers coexist: routers handle address translation and routing policies, while firewalls focus on security isolation. Extended products include Web Application Firewalls, secure gateways, and IDS/IPS.

Managing Congestion: Traffic Control Devices

Three main categories address congestion: Internet Behavior Management for granular traffic classification and control; Load Balancers / Application Delivery Controllers to distribute traffic across links and servers based on characteristics, applications, or addresses; Link Optimization devices at WAN edges to maximize low-speed link utilization.

Network Architecture Examples

Home SOHO Network

A typical home network uses a wireless router providing Wi-Fi access and routing to the external internet.

Small Business Network

Employs a two-layer, single-core topology with routers, switches, and servers.

Campus Network

Large enterprises and universities use a three-layer (access, aggregation, core) dual-core design. Zones (user, internal server, external server, management, internet) are interconnected and isolated via core switches and firewalls. Multiple internet egresses use routers for dial-up and NAT, traffic control devices for load balancing and behavior management, and firewalls for security isolation.

Data Center Network

A typical large Layer 2 data center / IDC design splits into tenant (service clusters), internet, and security management zones. The tenant zone uses device and link virtualization to boost processing and carrying capacity, with load balancers distributing traffic to servers. The internet egress zone runs BGP and address translation on routers, deploys IPS/anti-DDoS for volumetric attacks, uses traffic control for egress load balancing, and firewalls for isolation. The security management zone is accessed via firewall and monitored with audit, logging, IDS, and network management tools.

Source: Weak Current Documentation

Code example

--完--
读到这里说明你
喜欢
本公众号的文章，欢迎
置顶（标星）
本公众号 Linux技术迷，这样就可以第一时间获取推送了~
在
本公众号 Linux技术迷，后台回复：
Linux
，领取2T学习资料 ！
推
荐
阅
读
1.
Linux 中 find 命令的 35 个实际例子
2.
运维必备的《网络端口大全》，看这一份就够了
3.
Linux 学习指南 （收藏篇）
4.
2万字系统总结，带你实现Linux命令自由