Peer-to-Peer

In a P2P network every peer can request data and provide it. When you download a file with BitTorrent, released in 2001, you receive pieces from many other peers at once and upload pieces you already have to others. The more people share a file, the more capacity the network has, which is the opposite of a central server that gets slower as demand grows.

Peers still need to find each other. Some systems use a central tracker or directory just for discovery, as Napster did in 1999 before the transfer itself happened directly between users. Fully decentralized ones use distributed hash tables (DHTs), where peers share responsibility for looking up who has what. Because most devices sit behind NAT routers, P2P software also needs techniques for punching through to reach each other, with relay servers as a fallback.

Beyond file sharing, P2P ideas power blockchains such as Bitcoin, where every node keeps and verifies the ledger, WebRTC video calls that connect browsers directly, local multiplayer games, software update distribution and content-addressed storage such as IPFS.

A common misconception is that peer-to-peer means illegal file sharing. It is a neutral architecture with real strengths, no single point of failure, lower server costs and resistance to censorship, and real costs: harder security, since peers can be malicious, slower and less predictable discovery, and the difficulty of keeping shared data consistent.