How Does The Internet Work?

The internet is a global network of networks that connects billions of computers and devices. It enables browsing, email, streaming, and real‑time communication by coordinating layers of technologies that move data reliably and securely.

Overview

Network of networks interconnected by routers and ISPs
Data travels as small packets that may take different paths
Protocols define addressing, transport, security, and application behavior
DNS resolves human‑readable domains to IP addresses
The Web runs on HTTP; security is provided by TLS (HTTPS)
Performance is shaped by latency, bandwidth, and congestion

Packets, Routing, and Reliability

Applications break messages into packets; each packet carries source/destination IPs.
Routers forward packets using routing tables; paths can change due to congestion or failures.
Time To Live (TTL) prevents packets from circulating forever.
Reliability is handled above IP: TCP reorders, detects loss, and retransmits; UDP trades reliability for lower latency.
On the public internet, inter‑domain routing uses BGP so networks can choose economical and resilient paths.

IP Addresses and DNS

IPv4 (e.g., 192.0.2.172) and IPv6 (e.g., 2001:db8::1) identify devices; subnets and NAT organize local networks.
DNS flow: your resolver queries root → TLD → authoritative nameserver to get records.
Caching honors TTL to reduce latency and load.
Common records: A/AAAA map names to IPs, CNAME aliases names, MX routes mail, TXT carries metadata (e.g., SPF/verification).

Transport: TCP, UDP, and QUIC

TCP adds reliability via the three‑way handshake (SYN/SYN‑ACK/ACK), acknowledgments, sliding windows, and congestion control (e.g., slow start).
UDP is connectionless and used where minimal latency matters (streaming, real‑time).
QUIC (over UDP) provides transport‑level encryption and multiplexing; HTTP/3 uses QUIC to reduce head‑of‑line blocking.

HTTP and the Web

HTTP defines methods (GET, POST, etc.), status codes, headers, and bodies.
HTTP/1.1 uses persistent connections; HTTP/2 multiplexes streams over one connection; HTTP/3 runs over QUIC.
Typical page load:
- Resolve domain via DNS
- Connect/handshake (TCP or QUIC)
- Negotiate TLS for HTTPS
- Send request (method, path, headers)
- Receive response (status, headers, body)
Caching uses headers like Cache-Control, ETag, and Last-Modified to avoid redundant transfers.

HTTPS and TLS

HTTPS = HTTP over TLS for confidentiality, integrity, and server authentication.
Certificates chain from a leaf to intermediates to a trusted root CA; browsers verify the chain and hostname (via SNI).
Ephemeral key exchange enables perfect forward secrecy; OCSP/CRL check revocation; HSTS enforces HTTPS.

Physical Infrastructure

Bits travel over fiber, copper, and radio (Wi‑Fi/cellular); last‑mile connects homes/offices to ISP access networks.
Backbones and Internet Exchange Points (IXPs) interconnect ISPs; peering reduces cost and improves performance.
CDNs place content close to users to cut round‑trip times and offload origins.

What Developers Should Optimize

Timeouts, retries with exponential backoff, and idempotent operations.
Connection reuse, HTTP/2 or HTTP/3, and gzip/brotli compression.
Effective caching: far‑future caches for static assets, validators for dynamic.
DNS records and TTLs aligned with deploy strategy; use CNAMEs and health checks.
Observability: logs, metrics, tracing; measure latency, errors, and throughput.
Security hygiene: HTTPS everywhere, strong ciphers, certificate automation.

Learn More

Introduction: https://roadmap.sh/guides/what-is-internet
Developer guide: https://cs.fyi/guide/how-does-internet-work
MDN overview: https://developer.mozilla.org/en-US/docs/Learn/Common_questions/How_does_the_Internet_work
Short video: https://www.youtube.com/watch?v=7_LPdttKXPc