The invisible architecture that powers every click on the web
Every time you tap a link, a complex conversation begins. Behind the seamless interface of your browser lies the Hypertext Transfer Protocol, a sophisticated request-response system that has evolved from simple text exchanges into a high-speed, multi-layered engine of modern digital life.
Proposed in 1989 by Tim Berners-Lee, HTTP serves as the fundamental application layer for the World Wide Web [S2:p1, S2:p2]. At its core, the protocol operates on a client-server model: a user agent, such as a web browser, submits a request to a server, which then returns a response containing the requested resource, such as an HTML document [S1:p2, S1:p3]. While the protocol is stateless—meaning the server does not inherently retain user status between requests [S1:p39]—web applications maintain sessions using tools like cookies [S1:p39].
The protocol has undergone significant structural transformations to reduce latency. In the early HTTP/1.0 era, every single resource request required a separate TCP connection, which was highly inefficient [S1:p8, S1:p24]. HTTP/1.1 introduced persistent connections, allowing a single connection to be reused for multiple requests [S1:p9, S1:p25]. However, even with improvements like pipelining, issues like head-of-line blocking persisted [S1:p13, S1:p26].
Modern iterations have moved toward even greater efficiency. HTTP/2, released in 2015, transitioned from plaintext to a compressed binary representation of metadata and introduced multiplexing, allowing many concurrent requests through one connection [S1:p10, S1:p11, S1:p27, S2:p3]. The latest evolution, HTTP/3, abandons TCP in favor of the QUIC protocol over UDP [S1:p15, S1:p19, S1:p28]. This shift helps avoid congestion-related delays and can, in some cases, load pages over three times faster than HTTP/1.1 [S1:p17]. Today, the secure variant, HTTPS, protects the data of approximately 90.1% of websites [S1:p18].
Source: HTTP
