What Was the First Computer Network? The History of ARPANET Explained

ARPANET is commonly recognized as the first operational packet-switched computer network, though earlier systems make “first” harder to define. The answer depends on what counts as a network. Earlier projects linked computers or connected remote terminals, but they did not create the same kind of working network.

This history matters because research networks helped shape the internet. The network that began as a government research project later influenced how computers exchange information across the world. Its story also shows why one early milestone is not the same as the modern internet.

A look at the early claims, the technology behind them, and the network’s growth can clarify the difference. The first step is understanding what a computer network does.

Key Takeaways

• ARPANET is the usual answer when “first” means the first operational packet-switched network.
• Earlier systems connected computers or remote terminals for specific tasks.
• ARPANET began with four research sites in 1969.
• Shared protocols later helped separate networks communicate.
• The internet is larger than its early research network ancestor.

What Is a Computer Network?

A computer network is a group of connected devices that communicate and share information. In short, connected devices exchange information using agreed rules, and those rules help data reach the right place.

A node is a device or connection point that sends, receives, or forwards information on a network. A node can be a computer, a phone, or equipment that passes data between devices.

A computer network connects devices so they can exchange information using agreed rules.

For example, one computer can send a file to another over a home network. The computers follow shared communication rules, called protocols, so each can understand the exchange.

Networks can link a few devices in one room or connect systems across distant places. The basics of computer networks include how devices share data and resources. The internet is one large network of networks, not the only kind of network.

A private office network, for instance, can share printers without reaching the wider internet. A network can also connect computers across research sites, as later networks did.

This definition helps separate early remote links from the broader network ARPANET became. A single computer connection can be useful, but it does not always meet the same standard as an operational network.

What Was the First Computer Network?

The answer changes with the definition of “network.” By one measure, a remote computer link came first; by another, a military system predates ARPANET. ARPANET is the usual answer when the standard is an operational packet-switched network.

In 1940, George Stibitz linked a teletype at Dartmouth to a computer at Bell Labs for a real-time remote-computing demonstration. A user could send commands over the connection, but that milestone was not the same as an operational network linking many nodes.

The U.S. military’s SAGE radar system formed a computer network in the late 1950s. It linked radar sites and computers to support air defense, a different purpose and design from a research network.

So, what was the first computer network in the world? The answer depends on whether the term includes a remote computer link, a military system, or a general packet-switched network. These are separate milestones, not competing versions of one event.

ARPANET came online in 1969 and became a working packet-switched network. The ARPANET history records its development and early growth. Its goals explain why researchers built a network rather than another one-purpose connection.

Why Did ARPANET Begin?

J. C. R. Licklider imagined computers linked so people could reach information and programs from different locations. His early vision appeared in memos about an “Intergalactic Computer Network” in 1962 and 1963. He saw connected computers as tools for sharing information and computing power.

At the time, large computers often served one institution or a small group of users. Researchers wanted ways to reach distant systems without traveling to each site. The Advanced Research Projects Agency, or ARPA, backed research into that challenge.

Licklider’s vision of connected computers

Licklider’s idea was not simply to link machines. He pictured people using distant computer programs and information as part of their work. That vision helped frame the need for a network among research facilities.

Government research also needed dependable communication among sites in different places. Designers aimed to support information sharing while reducing reliance on one vulnerable connection.

Research goals and engineering barriers

The project had three practical needs:

• Remote access to computing resources at distant sites.
• Information sharing across research facilities.
• Communications that did not depend on one vulnerable connection.

Meeting those needs was difficult because computers and communication systems differed. A connection had to work across equipment that used different methods to store, process, and send information.

DARPA’s account of ARPANET and agency research describes how that work helped shape later network technology. Connecting sites was only one part of the problem. Designers also needed messages to travel without depending on one fixed route.

How Did Packet Switching Work?

Imagine sending one long message across a network whose routes may change. Packet switching breaks the message into smaller pieces that can travel through available links. The receiving computer puts those packets back in order.

A traditional phone call used circuit switching. The system reserved a path between callers for the length of the call. Packet switching instead shared network links among many messages.

Circuit switching versus packet switching

A reserved phone path worked well for a live call, but it could sit idle while one person paused. A shared network could use the same links for traffic from many computers.

Packet switching also let data take another route when a connection failed. It did not make every message secure, nor did it guarantee delivery on its own. Protocols had to help systems manage communication.

How a message travels in packets

A basic packet exchange follows four steps:

1. Divide the data into packets.
2. Address each packet for its destination.
3. Forward packets across available network links.
4. Reassemble the data at the receiving computer.

Packets from one message may travel along different routes. The receiving system uses information in the packets and communication rules to rebuild the message.

Packet switching was one of the key ideas behind ARPANET. An address for network communication helps identify where data should go. The first connected research sites needed equipment that could link each host computer to the new network.

Which Universities Joined ARPANET First?

The first four ARPANET nodes were UCLA, Stanford Research Institute, the University of California, Santa Barbara, and the University of Utah. Those four sites formed the network’s initial layout. Each site connected a host computer through an Interface Message Processor, or IMP.

An IMP was an early network connection device. It handled traffic between a host computer and the network, so each host did not need to manage every network link itself.

The sites were joined by 50-kilobit-per-second circuits. That speed was modest by modern standards, but it allowed computers at distant institutions to exchange data.

Stanford Research Institute was a separate institution from Stanford University. The similar names can cause confusion, but the original network node was in Menlo Park, not at Stanford University.

Each IMP helped manage the connection between its host computer and the network. The first test between two sites would show whether this arrangement could support a remote computer task.

What Happened During the First ARPANET Message?

“Lo” was the fragment that made the first ARPANET test memorable. The two letters crossed the network before the system stopped. The intended word was “login,” but the first attempt delivered only “lo.”

On October 29, 1969, Charley Kline sent a message from UCLA toward Stanford Research Institute. He was trying to log in to the computer at the receiving site, not start a public service or hold a complete conversation.

The exchange unfolded in three parts. Kline began typing “login,” the receiving system got “lo,” and the connection stopped before the word was complete.

The first message was recorded at 10:30 p.m. Pacific time. The system failure meant the test fell short of its intended task, but it still showed that computers could communicate across the new network.

That small success mattered because it confirmed the connection worked well enough to send data between distant computers. The next challenge was expanding the system beyond its first two communicating sites.

How Did ARPANET Grow Beyond Four Sites?

ARPANET grew from a four-site experiment into a network serving more research and government institutions. By 1970, it had nodes in several locations. Growth added users, computers, and new uses across the network.

Each new site brought different equipment and needs. Coordinating communication across varied computer systems became harder as the number of connected locations rose.

More institutions join the network

In 1970, ARPANET had nodes spread across several geographic locations. Over time, universities, government laboratories, and military bases joined. The wider network let more users share computing resources and communicate across sites.

A public demonstration brings attention

ARPANET’s first major public demonstration took place in Washington, D.C., in 1972. It showed what the network could do to an audience, but it did not open public access to the system.

By 1977, the network reached U.S. military bases, government laboratories, and universities. The mix of systems and users made ARPANET more useful, but it also raised a larger challenge: connecting separate networks.

When Did ARPANET Connect to Other Networks?

ARPANET’s first international connection came in 1973 through University College London. That link connected the U.S. network to British academic networks. It formed the first international heterogeneous computer network.

Heterogeneous means that the connected networks used different technologies or systems. They needed ways to exchange data even when their hardware and local communication rules did not match.

Internetworking requires several parts to work together:

• A physical or radio link between networks.
• Shared rules for addressing and forwarding data.
• Compatible ways to interpret the communication.

A new link alone cannot make two networks communicate. Each side must agree on how to identify destinations and handle information that crosses between systems.

The 1973 international connection did not instantly turn ARPANET into today’s global internet. It showed that separate networks could connect across national borders, while shared protocols would be needed to help them operate as an internet.

How Did TCP/IP Create the Internet?

As more networks connected, they needed a shared way to exchange data. TCP/IP provided a suite of communication protocols, or agreed rules. It is not one machine or another name for the World Wide Web.

Vint Cerf and Robert Kahn published a key paper on internetworking in 1974. The first TCP specification also appeared that year, setting out rules for communication across linked networks.

TCP helps manage communication

TCP, the Transmission Control Protocol, helps manage how data reaches a destination and comes back together. It can help computers detect missing data and put received pieces in order.

IP moves data between networks

IP, the Internet Protocol, helps direct data between networks. It uses addresses to help move packets toward their destinations. Together, TCP and IP helped different systems exchange information across network boundaries.

An initial TCP/IP implementation took place at Stanford in 1975. ARPANET switched to TCP/IP in January 1983, a major step in its evolution into the internet.

That change did not create every internet service at once. It gave separate networks a common way to communicate, making a network of networks more practical. People then began using services that ran over those connections.

How Did Email Change Network Use?

In 1988, Vint Cerf received permission to connect MCI Mail to the computer-based internet. The connection helped move networking beyond research and government users. MCI Mail began using it as a commercial email service in 1989.

Email let people send written messages without placing a live phone call. It also made it easier to contact people at other organizations linked through networks.

For users, email made three things possible:

• Send written messages to another person.
• Reach people across connected organizations.
• Communicate without a live phone conversation.

Network email could reach beyond one local system as connections expanded. A message could cross between services without requiring both people to be online at the same time.

Connecting a commercial email service did not mean every business or household had internet access. Commercial messaging was an early step toward broader use, while household access spread more slowly.

When Did the Internet Reach Homes?

U.S. households began gaining internet access through commercial dial-up services during the 1990s. There was no single start date for everyone. Home access spread over time and expanded further through the 2000s.

A household could connect a computer to an internet service provider through a telephone line. The modem used that line to send data, and the line was usually unavailable for regular phone calls during the session.

For people asking when did the internet start in homes, the answer is a gradual shift, not one launch day. Access depended on services reaching an area and households choosing to connect.

Dial-up made the internet available from home, but it could feel different from a modern connection. A computer had to dial in, and the phone line stayed tied up while the user was online.

Household use grew over the 1990s and 2000s, though access never began for everyone at once. Research backbones and new information tools also helped support wider use.

What Did NSFNET and the World Wide Web Add?

NSFNET connected research and education sites, while the World Wide Web made linked information easier to access. They served different roles: one was a network backbone, and the other was an information system. The internet carries data; the Web is accessed through the internet.

The National Science Foundation launched NSFNET in 1986 as a research and education network. It linked academic sites and supported wider networking beyond earlier research systems.

Tim Berners-Lee developed the World Wide Web at CERN. The Web used HTML to create pages, URLs to name resources, and HTTP to move information between computers.

Those technologies helped people find and follow linked information. The Web was not the internet itself; it was a service that used internet connections.

The National Science Foundation shut down its dedicated backbone in 1995 as commercial internet services expanded. That change marked the growth of commercial networks, not the end of the internet.

Research and commercial systems grew into different network types. Knowing their reach helps explain how a small local network differs from one linking distant sites.

What Are the Main Types of Computer Network?

The types of computer network are often grouped by the area they connect. Some networks stay local, while others link sites across cities or countries. Reach is a useful way to tell these network types apart.

A local area network, or LAN, serves a home, school, or office. A metropolitan area network, or MAN, can connect locations across a city. A wide area network, or WAN, links sites over much greater distances.

Local networks connect nearby devices

A LAN can connect computers, printers, and other devices within one building. The devices may share files or a printer, but a LAN does not have to connect to the internet.

Wide-area networks link distant sites

WANs connect systems in different places. ARPANET linked distant research sites, while the internet joins many networks around the world.

These labels describe reach, not a single set of equipment. Different network types may use different hardware and protocols, based on their needs and design.

Knowing a network’s reach helps people understand what it can connect. It also points to the benefits and tradeoffs that come with linking devices.

What Advantages and Risks Do Networks Bring?

Sharing information across devices can save time, but it can expose data to people who should not see it. That tradeoff is part of using connected systems. Connectivity by itself does not guarantee privacy or security.

The advantages of a computer network include faster communication, access to shared information, and shared services. A workplace network, for example, can let staff use the same files or printer.

Networks also create risks. Unauthorized users may gain access, personal data may be exposed, and people may depend on connected services that can fail.

Useful habits can lower some risks:

• Use strong authentication for accounts and devices.
• Share only the information that others need.
• Keep devices and software updated.

These steps are not a complete defense. A strong password cannot fix every flaw in a system, and careful sharing cannot control every person who gets access.

Networks make information easier to share, but they also make control and security important. These tradeoffs connect to design ideas that early networks helped establish.

How Does ARPANET Shape Networks Today?

A mobile device can pass information to another device even when fixed infrastructure is unavailable. That is one example of mesh networking, where devices can relay data through nearby nodes. Today’s networks carry forward ideas such as packet-based communication and shared protocols.

Mesh networks can help devices communicate without relying on one fixed connection point. DARPA-supported research has explored mobile mesh concepts for communication among devices and connected systems.

Modern networks carry forward early ideas, but they are far larger and use technologies that did not exist on ARPANET.

The Internet of Things, or IoT, includes connected devices that exchange data through networks. Sensors, machines, and other devices can send information without using a person’s computer as the main link.

These systems reflect the value of connecting separate devices and networks through shared rules. They also use modern services and technologies that were not part of the original research network.

ARPANET is a historic foundation, not today’s internet. The modern internet is far larger, with many types of networks and services built since those early links.

Conclusion

ARPANET’s historic role differs from the modern internet. It is commonly identified as the first operational packet-switched network, while the internet developed as many networks learned to communicate.

One useful next step is to trace a message from a device, through a router, to an online service. At each point, note where shared protocols help the data move.

That distinction makes it easier to use the terms network, internet, and Web with care. Each names a different part of how connected systems work.