The history of the internet: from ARPANET to today's global network

The internet is not the same thing as the web

Before diving into the history, it's worth clarifying terminology that everyday speech tends to blur. The internet is the global infrastructure of interconnected networks — physical cables, routers, and protocols (covered in detail in our article on TCP/IP) that let any device send data to any other. The World Wide Web is just one of many services that run on top of this infrastructure — a system of interlinked documents accessible over HTTP. Email, DNS, and internet telephony all run on that same internet too, but the web itself only emerged two decades later.

The Cold War and an idea for a network that wouldn't die

The basic idea behind the internet emerged independently in two places in the early 1960s. In the US, Paul Baran at the RAND Corporation was working on the problem of how a communications network could survive partial destruction — a nuclear strike on one of its nodes, for example. His answer was a decentralized network with no single central point, in which data travels in small, independently addressed blocks that can find an alternate path on their own if part of the network becomes unavailable.

Independently, Donald Davies at Britain's National Physical Laboratory arrived at a virtually identical conclusion — and it was he who coined the term "packet" for these small units of data. This idea stood in direct contrast to how the telephone network operated at the time.

This property — resilience against the loss of individual nodes — later became part of the popular myth that the internet was built specifically to "survive a nuclear war." The reality is subtler: resilience was an important design property, but the direct impetus for funding the project that became the internet was, first and foremost, the desire to efficiently share expensive, scarce computing resources among distant research institutions.

ARPANET: the first four computers (1969)

The American agency ARPA (later DARPA) funded the construction of an experimental network to test these theories — ARPANET. Connecting the individual computers was handled by specialized devices called IMPs (Interface Message Processors), the ancestors of today's routers, built by BBN Technologies.

The first connection attempt took place on October 29, 1969, between the University of California, Los Angeles (UCLA) and the Stanford Research Institute (SRI). An operator at UCLA tried to log in remotely by typing "LOGIN" — the system on the other end crashed right after the letters "L" and "O" got through. An hour later, the connection was successfully completed. By the end of 1969, two more nodes had joined — the University of California, Santa Barbara (UCSB) and the University of Utah — creating the first network with four nodes.

The first protocols and the "killer app"

Computers on ARPANET initially communicated via the NCP protocol (Network Control Protocol) — simple, but with limitations that only became apparent as the network grew larger. Surprisingly, the most-used application on this military-funded research network turned out to be something entirely civilian: email. In 1971, Ray Tomlinson programmed a system for sending messages between different computers on the network and chose the @ symbol to separate the username from the computer name — a choice that has survived to this day in almost unchanged form.

TCP/IP: the protocol that unified the networks (1974 – 1983)

As more, mutually different networks appeared (not just ARPANET, but also radio and satellite networks), the need arose for a universal way to connect them into a "network of networks." Vint Cerf and Bob Kahn solved this in 1974 with the design of the TCP/IP protocol, covered in detail in our dedicated article. The key property of the design was that individual networks could keep their own internal technology — TCP/IP acted as a shared "language" at the boundaries between them.

The switch from NCP to TCP/IP wasn't gradual — it happened all at once, on January 1, 1983, a day still jokingly called "flag day": every computer on the network had to switch to the new protocol simultaneously, or it would stop being able to communicate with the rest of the network. This date is considered the symbolic birth of the internet as we know it today — a network of networks joined by one shared protocol.

The growth of local networks and academic infrastructure (the 1970s–80s)

• Ethernet (1973) — Bob Metcalfe at the Xerox PARC research center developed a technology for connecting computers within a single building or campus, which became the universal standard for local area networks (LANs) and remains the basis of today's home and corporate networks.
• USENET (1979) — a decentralized system of discussion groups created by Tom Truscott and Jim Ellis, the ancestor of today's internet forums, connecting universities that weren't yet part of ARPANET.
• BITNET and CSNET (early 1980s) — additional academic networks extending network access to universities outside the original ARPANET circle.
• NSFNET (1985 – 1986) — the US National Science Foundation built a high-speed backbone network connecting regional academic networks. It gradually took over as the main backbone from the original, by then outdated ARPANET, which was finally decommissioned in 1990.

DNS: why you don't have to remember numbers (1983)

While the network was small, the list of all computers and their numeric addresses was maintained in a single text file, HOSTS.TXT, centrally distributed to every connected computer. As the number of nodes grew, this approach became unsustainable. The solution came in 1983, when Paul Mockapetris proposed DNS (Domain Name System) — a decentralized, hierarchical system that can translate human-readable names (e.g. cyberforce.sk) into numeric IP addresses without needing a single central list. DNS remains one of the cornerstones of the internet to this day, and we cover its principles in detail in our article on TCP/IP.

The World Wide Web: the internet gets a face (1989 – 1991)

Until 1989, the internet was a tool used almost exclusively by scientists and military researchers — with no graphical interface, accessible only through the command line. That changed thanks to British scientist Tim Berners-Lee, working at the CERN research center in Switzerland. In March 1989, he submitted a proposal titled "Information Management: A Proposal," describing a system of interlinked documents accessible over the network — a solution to the problem of how CERN scientists could easily share and find information.

By the end of 1990, Berners-Lee had a working prototype of the three technologies that still form the foundation of the web today:

• HTML (HyperText Markup Language) — a language for formatting and linking documents.
• HTTP (HyperText Transfer Protocol) — a protocol for transferring those documents between a server and a browser.
• URL (Uniform Resource Locator) — a unified way to address a specific document anywhere on the network.

Berners-Lee launched the first web server and the first web browser (simply named "WorldWideWeb") on his NeXT computer at CERN in late 1990. The web became publicly available outside CERN in 1991, and in 1993 CERN decided to release the web's underlying technology completely free of charge, with no licensing fees — a decision now widely regarded as one of the key reasons the web was able to spread across the world so quickly and freely.

Browsers and the "browser wars" (1993 – 1998)

Genuine mass public interest wasn't ignited until the graphical browser Mosaic (1993), developed by a team led by Marc Andreessen at the University of Illinois — the first browser able to display images directly inline with text, rather than only in separate windows. Andreessen went on to co-found Netscape, whose browser, Netscape Navigator (1994), became the dominant way ordinary people accessed the internet for a few years.

Microsoft responded by releasing its own browser, Internet Explorer, which was gradually integrated directly into Windows — kicking off the era of the so-called browser wars, a competitive battle for dominance of the browser market that had a fundamental impact on how the web evolved as a platform over the following decades.

Commercialization, the dot-com era, and its collapse (1994 – 2001)

By the mid-1990s, the internet had fully opened up to commercial use. Millions of households connected via slow but revolutionary dial-up connections, often through providers like AOL. As the amount of content grew, so did the need for navigation — the first directories and search engines appeared, like Yahoo! (1994) and AltaVista, until Google arrived on the scene in 1998 with a significantly more relevant results-ranking algorithm, becoming the world's dominant search engine within a few years.

Easy access to venture capital and enthusiasm around internet companies fueled the so-called dot-com bubble — a period when investors funded hundreds of internet startups regardless of their actual profitability. The bubble burst in 2000 – 2001, causing a deep stock market crash and the collapse of a large number of companies — but the infrastructure and underlying technologies this boom built survived, and became the foundation for the next, this time more durable, phase of growth.

Broadband, Web 2.0, and the mobile era (2000 – 2010)

• Broadband connections gradually replaced slow dial-up lines, enabling richer content — video, larger images, interactive applications.
• Wikipedia (2001) demonstrated the power of content created collectively by users, rather than a central editorial staff.
• Social networks — Friendster and MySpace paved the way, but it was Facebook (publicly available from 2006) and later other platforms that definitively turned the web from a place for reading into a place for sharing and interaction — a period collectively referred to as Web 2.0.
• YouTube (2005) made video uploading and sharing accessible to the masses, something that had previously required specialized technical knowledge.
• The iPhone (2007) and the ensuing wave of smartphones with mobile internet (3G, later 4G) moved the internet's center of gravity from the desktop computer to every user's pocket — today, the vast majority of internet access worldwide comes from mobile devices.

Today's internet: cloud, IoT, and constant growth

Today's internet is disproportionately more complex than the original experimental network of four computers. Key trends of the last decade include cloud computing (compute and storage as a service, rather than owned hardware), massive CDN networks (Content Delivery Networks) bringing content geographically closer to users, an explosion of IoT devices (the Internet of Things) — from smart homes to industrial sensors — and the gradual shift from the exhausted IPv4 address space to IPv6, which we covered in our TCP/IP article. A network originally designed for a few dozen academic computers today connects tens of billions of devices.

Key milestones

Security grew alongside the network — from the first worm to today

Threats showed up surprisingly early in the internet's history. In 1988, student Robert Tappan Morris released the so-called Morris Worm — the first internet worm, originally intended only to quietly spread and measure the size of the network, but a bug in its code caused computers to get infected repeatedly and overload themselves to the point of collapse. It's estimated to have knocked out tens of percent of then-connected computers, and led to the founding of the first CERT/CC (Computer Emergency Response Team) — an institution that still coordinates responses to large-scale security incidents today.

Since then, security threats have grown hand in hand with the growth and commercialization of the network — from early viruses, through the spam of the AOL era, DDoS attacks targeting dot-com companies, all the way to today's sophisticated threats described in our article on the most common cyberattacks. The reverse is also true: the more the internet is integrated into critical infrastructure, the economy, and everyday life, the higher the stakes when its security fails.

Summary

From the theoretical design of a resilient, decentralized network at the RAND Corporation, through the first shaky connection between UCLA and SRI in October 1969, the unifying TCP/IP protocol of 1983, all the way to Tim Berners-Lee's web, which gave the internet a face people could understand — the history of the internet is a story of simple, robust ideas layered gradually on top of one another. No single inventor or company "created" the internet in one moment; it emerged as the sum of dozens of partial solutions over nearly sixty years, and its development continues today at exactly the same pace — building new layers on top of an already existing foundation.