NSFNET

Considering how much of modern life revolves around high-speed global connectivity — from remote work and online shopping to social media and endless sources of entertainment — it’s almost difficult to imagine that a high-functioning society could have ever existed without it. But the internet had to be invented, and as has so often been the case with foundational technologies, IBM played a significant role in both its development and proliferation.

Its creation stems in part from the desire of scientists to collaborate. Prior to the early 1980s, this typically meant traveling, often around the globe, to compare research and share computing resources. Around that time, an early regional telecommunications network, spearheaded by the Advanced Research Projects Agency and funded by the Pentagon, splintered into separate military and civilian networks, known respectively as MILNET and ARPANET. ARPANET was complicated, slow and accessible only by highly trained operators at academic institutions. It was based on a now-defunct communications protocol and had little ability to connect to other networks. Even so, by linking several national laboratories and supercomputing centers, it streamlined scientific collaboration and hinted at the transformative potential of a global network.

Not long after, in 1985, the National Science Foundation (NSF) launched an initiative to build a more powerful, flexible and inclusive backbone to link supercomputer centers and regional academic networks based on the novel Transmission Control Protocol/Internet Protocol (TCP/IP). A year later, the agency unveiled NSFNET. It was slow (supporting transmission speeds of 56 kilobits per second) and continuously overloaded, but it exhibited obvious potential and so the agency solicited proposals to build and maintain a higher-speed version. In late 1987, the job went to a team comprising IBM, the telecom company MCI, and Merit, a not-for-profit networking organization with members from Michigan universities.

Their efforts would give rise to the internet — and forever change life on our planet.

According to NSFNET co-principal investigator Hans-Werner Braun, the idea of creating a fast network of networks based on a standard communications protocol had immediate and universal appeal — but there was also a fair amount of skepticism that such a project was even technologically feasible. The view inside IBM was more optimistic.

The company quickly assembled a team from across its ranks and created an environment at the project’s Michigan headquarters that was part assembly line, part skunkworks. Computers and peripherals flowed in for testing and configuration before being redirected to various campuses and supercomputer sites. “We had the whole floor covered with parts and machines and boxes; it was a great way to deploy everything,” said Elise Gerich, the site liaison.

Al Weis and Barry Appleman from IBM Research, and Bob Mazza, Walter Wiebe and Rick Boivie from IBM’s Academic Information Systems division led hardware, software and project management, overseeing a development process that resembled a game of leapfrog. One team would invent a new class of hardware to boost transmission speeds — which would then cause problems with data consistency, forcing another team to create software solutions. And so it went: engineers solving one problem only to reveal another.

Some IBMers logged 100-hour weeks on this routine for months at a time. IBM’s project manager, Harvey Fraser, recalled that “the team worked well because we had resources, executive time and the desire to make it a success.”

In July 1988, the network went live, linking 170 academic and research communities together via a high-speed, reliable data network service spanning the United States. It worked so well that ARPANET was soon shut down. “We met all the deadlines. And it created a big buzz in the university academic community and in the supercomputing community,” said John A. Armstrong, who was director of the IBM Thomas J. Watson Research Center in the late 1980s.

Traffic soared almost immediately. In the first year, only users in the US, France and Canada could access the network. Between 1989 and 1993, 10 to 12 additional countries were added each year, and 21 were added in 1994. The pace of uptake required the technology partners to continue developing innovative ways to expand capacity and increase speed. The IBM/MCI/Merit team upgraded the capacity of network links to 1.5 Mbps and then to 45 Mbps.

From the project’s inception, IBM leadership was convinced that NSFNET would be important to the nation, but they also felt it could unveil new pathways for the company. “IBM was not a player in the world of supercomputing at the time in the ’70s and early ’80s,” Armstrong said. “And so an opportunity came along for IBM to get back into the supercomputing business, very indirectly.”

The implications extended much further. The NSFNET backbone was decommissioned in April 1995, giving way to a fast-emerging, commercially oriented internet. This marked a significant transition for society and for IBM alike. The company began systematically building new networking capabilities into its software products, a process that led to the development of internet-related software and consulting offerings over the next 20 years. It introduced OS2/Warp, the first operating system to offer TCP/IP connectivity, email, IBM Global Network access and IBM’s WebExplorer browser. It released Bamba, a streaming audio technology and, together with Sears, one of the first commercial internet service providers, Prodigy. More important, IBM’s work on NSFNET instigated the company’s forays into developing, deploying and servicing technologies and solutions that would become foundations for critical revenue streams: networked computing, network security and e-business.

The rest of the story is by now highly familiar. Every day, scientific researchers around the world seamlessly collaborate on world-changing projects and ideas. Businesses rely on the internet to communicate with customers and deliver products and services. And the general public’s appetite for instantaneous communication, endless streams of music, movies and boundless shopping grows more voracious by the day. In the 1970s, the modern internet would surely have seemed like sci-fi fantasy. But a relatively small group of committed technologists — many of whom were IBMers — made it real. And the world will never be the same.