This year marks the 30th anniversary of the fully commercial internet

The internet began as a US government project to connect academic researchers together. How did it shift into the commercial phenomenon that we know today? And is the US government still involved?

Thirty years ago, on 30 April 1995, the modern internet as we know it today began. This was the day the US government’s National Science Foundation Network (NSFNet), which had connected academic and research sites across the country, was decommissioned. In its place, commercial forces drove the development of the internet, resulting in exponential growth, investment and innovation.

From academic intent…

NSFNet was developed to enable academic and research access to five supercomputers across the USA. It was a true backbone in the sense that all the smaller research networks connected to it, and it carried their traffic across the country to and from every other network. When it started in 1986, it connected six research centres and operated at a speed of 56 kilobits per second (kbit/s) – the top speed of a dial-up modem. It was then upgraded in 1988 to operate at 1.5 megabits per second (Mbit/s) with 13 nodes, or network hubs. By 1991, NSFNet had evolved, increasing its capacity to 45Mbit/s and linking 16 hubs across the nation.

…to commercial involvement 

Due to its research purpose, NSFNet was provided free of charge to the academic and research networks. It had an acceptable use policy preventing commercial traffic from taking advantage of access. Over time, as the use of the internet grew, commercial internet service providers (ISPs) began to emerge, and the acceptable use policy started to be relaxed. Eventually, it became clear that the commercial internet could replace the role played by NSFNet, which was decommissioned after a period of transition.

A change in infrastructure…

A direct legacy of the transition was the creation of interconnection arrangements that remain in place to this day. To replace the NSFNet role as the hub of interconnection, the NSF designated four network access points (NAPs) across the USA. These NAPs were designed to allow any network to connect and exchange traffic with others. They soon became congested, and new internet exchange points (IXPs) were developed. These enabled efficient traffic exchange within a country, region or city, rather than across borders or oceans. Today, almost every country has at least one IXP, with some being for profit and many others being non profit and run by their members.

In place of the NSFNet, which served as a backbone for exchanging traffic between connected networks, commercial ISPs developed peering and transit arrangements. In a peering arrangement, two networks agree to exchange each other’s traffic for delivery, often without payments, if they consider themselves ‘peers’ based on size or traffic. In a transit arrangement, one network pays another to deliver its internet traffic to and from the rest of the internet. These arrangements have created the network of networks that is the internet, currently consisting of at least 90 000 independent interconnected networks.¹ Despite the changes since the commercialisation of the internet, these arrangements are still in use today.

…and a changed landscape

The pace of change in the internet has been impressive. At the end of 1994, there were 20 million internet users, many still in the USA. Today there are over 5.5 billion users worldwide, although the remaining digital divide is still too large. At the end of 1994, there were just 2738 websites. The Mosaic browser was developed at one of the NSFNet supercomputer sites in Illinois and was introduced only in 1993, and the first commercial browser, Netscape Navigator, launched at the end of 1994. Today there are 1.1 billion websites, along with almost 9 million mobile apps.

A trillion-dollar industry

With this explosive growth came significant innovation and investment that drove the commercialisation of the internet and its globalisation. New companies such as Google (founded in 1998) and Facebook (2004) emerged, reshaping the digital landscape. Meanwhile, established giants like Apple and Microsoft pivoted from their roots in hardware and software to offering online services and devices. 

Additionally, new types of content such as social media and video streaming emerged, along with new business models like content delivery networks that distribute much of this content. Significant investments have been made to support this distribution, with hundreds of billions spent on submarine cables, terrestrial connectivity, and fixed and mobile broadband networks. 

The decision to leave the internet to commercial forces was forward thinking. While it recognised the increasing growth of commercial networks, it also clearly helped to propel that growth. Instead of relying on NSFNet to deliver and exchange traffic, companies were forced to invest further in their own networks and develop interconnection points and arrangements to do so. The US government’s strategy of initially developing the backbone while maintaining control, and then relinquishing that control, mirrored its approach with the internet. This foresight led to the internet as we know it today.

¹ Each network has a unique autonomous system number (ASN). While more than 100 000 have been allocated to networks, estimates are that around 90 000 are being used.