The History of the Internet: From ARPANET to the Modern Web
How packet switching, ARPANET, TCP/IP, DNS and the World Wide Web transformed experimental computer links into global communications infrastructure.

The internet did not spring from a single invention or one decisive breakthrough. It developed over decades as researchers found new ways for computers to share resources, exchange data and communicate across otherwise incompatible networks. Government funding, university research, open standards and private investment all contributed to the infrastructure now used around the world.
Understanding that history requires an important distinction: the internet and the World Wide Web are not the same thing. The internet is the global system of interconnected networks that carries data. The Web is one service that operates on that infrastructure, using linked documents and applications. Email, file transfers, online games and internet-based voice calls do not have to be part of the Web.
Before the internet: Expensive computers and shared access
Computers of the 1950s and early 1960s were large, costly machines generally owned by governments, universities and major organizations. Because providing a separate computer to every user was impractical, institutions created time-sharing systems that let multiple people work with a central machine.
Researchers soon looked beyond sharing one local computer. They began considering whether people could access programs, processing power and information stored on machines in other locations. In 1960, psychologist and computer scientist J.C.R. Licklider outlined ideas that anticipated interactive computing and a globally connected information environment.
Meanwhile, researchers at MIT, RAND and Britain’s National Physical Laboratory independently explored packet switching. Conventional circuit-switched communication reserves a continuous connection for the duration of a session. Packet switching instead breaks information into smaller units that can move through a network and be reassembled at the destination. This approach uses communication links more flexibly and efficiently.
The internet’s origins are sometimes reduced to a defense project intended to survive nuclear war. Defense support was significant, but that explanation is incomplete. Resource sharing, interactive computing and communication among research institutions were also central goals.
ARPANET puts packet switching to work
The Advanced Research Projects Agency Network, better known as ARPANET, transformed packet-switching research into an operating network. Its first Interface Message Processor was installed at the University of California, Los Angeles, in September 1969.
On October 29, 1969, a UCLA computer attempted to log in to a machine at Stanford Research Institute. The connection failed after sending the first two letters of the command, making LO the first host-to-host message transmitted over ARPANET. The complete LOGIN command was successfully sent later that evening.
By the end of 1969, ARPANET connected four institutions:
- UCLA
- Stanford Research Institute
- UC Santa Barbara
- The University of Utah
Although modest in scale, this network demonstrated that computers separated by long distances could communicate and share resources through an operational packet-switched system.
Email changes the purpose of networking
Remote login and file transfer were important early network applications, but email helped turn networking into a social medium. In 1972, Ray Tomlinson developed software for sending messages across a network and adopted the @ symbol to separate a recipient’s user name from the host computer. Larry Roberts later added features for listing, reading, filing, forwarding and replying to messages.
Email quickly became one of ARPANET’s most influential uses. It demonstrated that networks could connect people as effectively as they connected machines. Mailing lists and later systems such as Usenet expanded that model by supporting discussions and online communities.
TCP/IP creates a network of networks
ARPANET’s original communication protocol was intended for hosts within one network. A broader system needed to connect independent networks without requiring them to use identical internal designs.
During the 1970s, Robert Kahn and Vint Cerf developed an architecture that evolved into the Transmission Control Protocol and Internet Protocol, collectively called TCP/IP. The two protocols perform complementary roles:
- IP provides addressing and routes packets between networks.
- TCP supports reliable delivery when applications require it, identifying missing data and restoring packets to the proper order.
The architecture was intentionally general. Individual networks could remain autonomous, routers could carry traffic between them and ordinary operation did not depend on a single global controller.
On January 1, 1983, every ARPANET host was required to migrate from the older Network Control Protocol, or NCP, to TCP/IP. This coordinated change, sometimes described as a protocol “flag day,” is widely regarded as a defining step toward the modern internet. A common protocol suite could now support continued growth across many different networks.
DNS replaces difficult numeric addressing
Early users depended on a centrally maintained file that associated host names with numeric network addresses. As more computers connected, updating and distributing that file became increasingly impractical.
The Domain Name System, designed in the early 1980s, distributed name resolution across a hierarchy of servers. DNS allowed people to use memorable names rather than typing IP addresses directly. It also provided a scalable structure for domains, including .com, .org, .edu and country-code domains.
DNS usually operates out of sight, but it remains a fundamental part of internet infrastructure by translating human-readable names into the addresses computers use to locate one another.
NSFNET carries the internet beyond ARPANET
Personal computers, local area networks and academic networking expanded rapidly during the 1980s. The US National Science Foundation’s NSFNET connected higher-education and research communities while requiring TCP/IP, helping the protocol suite spread far beyond ARPANET.
ARPANET was decommissioned in 1990, but the larger internet continued through the protocols, operational knowledge and connected networks the project had helped establish. As restrictions on commercial activity eased and private backbone providers expanded, the internet began moving from a research-oriented system toward a public and commercial medium.
The NSFNET backbone was retired in 1995, by which time private infrastructure had developed enough to assume its role.
The World Wide Web emerges at CERN
The internet had already existed for years before the first website appeared. In 1989, British computer scientist Tim Berners-Lee, then working at CERN, proposed a way to connect and retrieve information across the organization.
By the end of 1990, he had developed the Web’s essential components:
- HTML for structuring documents
- HTTP for transferring those documents and related resources
- URLs for identifying where resources could be found
Berners-Lee also created the first web browser, web server and website. Hosted on his NeXT computer, the initial site described the Web project and explained how to build pages and servers.
On April 30, 1993, CERN placed the Web software in the public domain and later issued an open release. Removing proprietary licensing barriers allowed developers and organizations to implement the technology freely, accelerating its adoption.
Graphical browsers open the Web to the public
Early browsers established that the Web worked, but Mosaic made it more approachable. Released in 1993, the graphical browser presented text, images and links through an interface that a much broader audience could navigate. Netscape Navigator followed in 1994.
The late 1990s brought search engines, web portals, online stores and intense investment in internet companies. The dot-com crash of 2000 eliminated many businesses, but it did not stop the underlying expansion of internet access. Broadband gradually replaced slow dial-up service in many homes, while browsers and web standards supported increasingly interactive applications.
Participation, smartphones and cloud computing
During the 2000s, creating and publishing online content became much easier. Blogs, Wikipedia, social networks, video platforms and collaborative tools encouraged people to contribute rather than simply read. The label Web 2.0 came to describe this participatory period, although it did not represent one specific technical upgrade to the Web.
Smartphones then changed both the location and frequency of internet use. Mobile applications, Wi-Fi and cellular broadband made connectivity an everyday, often continuous experience. At the same time, cloud computing shifted large amounts of processing and storage into data centers. Services could operate at global scale while users reached them through browsers and mobile apps.
The internet remains a work in progress
The modern internet carries entertainment, commerce, education, government services and private communication. Its physical foundation includes fiber-optic cables, terrestrial and mobile networks, satellites, routers, internet exchange points and vast data centers. Open protocols allow equipment and services operated by different organizations to work together.
Technical openness does not make every part of the system equally decentralized. A relatively small group of platforms, cloud providers and network operators processes a substantial share of online activity. The internet also continues to face unresolved challenges, including:
- Unequal access to reliable connectivity
- Surveillance and privacy risks
- Cybercrime and network security threats
- Misinformation
- Market concentration
- Energy consumption
- The slow transition from IPv4 to IPv6
The central lesson of internet history is that the network is not a completed invention. It is an evolving arrangement among independent networks, standards organizations, governments, researchers, businesses and users. ARPANET provided an experimental base, TCP/IP made broad interconnection possible, DNS made destinations manageable and the Web brought accessible publishing and navigation to a mass audience. Together, those developments transformed a four-node research network into global infrastructure.
How this article was prepared
Uses consistent metric definitions, excludes invalid samples where identifiable, compares medians rather than isolated extremes, and describes device, access-network, and geographic limitations. Writing approach: Measurement-led troubleshooting starts with an observable symptom and a precisely defined metric. Controlled comparisons change one variable at a time. Results, possible causes and the evidence needed to confirm them remain clearly separated.
Read our methodology →Reviewed by the Internet Analysis Editorial Team
Reviewed by the Internet Analysis Editorial Team · Updated August 15, 2026
Meet the editorial team →Article context, review and related questions
How packet switching, ARPANET, TCP/IP, DNS and the World Wide Web transformed experimental computer links into global communications infrastructure.
| Measure | Value | Context |
|---|---|---|
| Article type | Technology History | Editorial classification |
| Reading time | 7 minutes | Estimated at approximately 220 words per minute |
| Editorial review | Internet Analysis Editorial Team | Updated August 15, 2026 |
| Review date | August 15, 2026 | Latest stored article update |
Methodology
Uses consistent metric definitions, excludes invalid samples where identifiable, compares medians rather than isolated extremes, and describes device, access-network, and geographic limitations. Writing approach: Measurement-led troubleshooting starts with an observable symptom and a precisely defined metric. Controlled comparisons change one variable at a time. Results, possible causes and the evidence needed to confirm them remain clearly separated.
Full methodology →Data freshness
- Page updated
- Data period
- August 15, 2026
- Responsible editor
- PiotrNetwork Performance Analyst
Primary sources
- Internet Analysis editorial methodologyReview and limitation rules
Limitations
- The article is informational and may simplify technical details for readability.
- Products, standards, prices and service availability can change after the review date.
- The latest review date does not guarantee that every external product or service remains unchanged.
Related questions
What is the main point of “The History of the Internet: From ARPANET to the Modern Web”?
How packet switching, ARPANET, TCP/IP, DNS and the World Wide Web transformed experimental computer links into global communications infrastructure.
How was this article prepared?
Uses consistent metric definitions, excludes invalid samples where identifiable, compares medians rather than isolated extremes, and describes device, access-network, and geographic limitations. Writing approach: Measurement-led troubleshooting starts with an observable symptom and a precisely defined metric. Controlled comparisons change one variable at a time. Results, possible causes and the evidence needed to confirm them remain clearly separated.
When was this information last reviewed?
The latest stored review or update date is August 15, 2026.
