In the early 1970s, Larry Roberts approached AT&T, the vast American telecommunications monopoly, with an intriguing offer. At the time, Roberts was director of the computing division of the Advanced Research Projects Agency (ARPA), a relatively young organization within the Department of Defense that was dedicated to long-term, blue-sky research. Over the previous five years, Roberts had overseen the creation of ARPANET, the first computer network of any significance, which now linked computers at about twenty-five sites across the country.
The network was a success, but its long-term operation, and all of the bureaucratic work which that entailed, did not fall within ARPA’s mandate. Roberts was now looking to offload the task to someone else. And so, he contacted executives at AT&T to offer them the keys to the system.1 After mulling the matter over, AT&T ultimately rejected the offer. Its business and engineering leadership believed the fundamental technology upon which ARPANET operated was impractical and unstable, and had no place in a system designed for reliable, universal service.
ARPANET, of course, was the seed around which the Internet crystallized, the prototype of a vast, world-circling information system, whose kaleidoscopic capabilities defy enumeration. How could AT&T have been so blind to this potential, so mired in the past, we are led to wonder? Bob Taylor, who had recruited Roberts to oversee the building of ARPANET back in 1966, later put it quite bluntly: “Working with AT&T would be like working with Cro-Magnon man,” he recalled.2 However, before we raise our hackles too sharply at the brutish ignorance of these anonymous corporate bureaucrats, let us step back a moment. The history of the Internet will be the subject of our story, so it would be good to first take a broad view of what it is we are talking about.
Of all the technological systems constructed in the latter half of the twentieth century, the Internet is surely of the most profound significance to the society, culture, and economy of our contemporary world. Perhaps jet travel is its closest rival. Using the Internet, individuals can instantly share pictures, videos, and thoughts, welcome or unwelcome, with friends and family across the world. Young people a thousand miles apart now regularly fall in love, and even marry, within the confines of a virtual world. A never-ending shopping mall is instantly accessible at all hours of day and night from the comfort of millions of homes.
Most of this is trite and familiar enough. But, as the author can attest, the Internet has also proved perhaps the greatest distracter, time-waster, and general source of mind-rot in human history, surpassing television – no mean feat.3 It has made it possible for fanatics, zealots and conspiracy theorists to spread their nonsense across the globe at the speed of light – some of it harmless, some much less so. It has enabled many organizations, public and private, to slowly amass, and, in some cases, quickly and embarrassingly lose, draconic hoards of data.4 It is, in short, a vast amplifier of human wisdom and folly, and the latter is dismayingly abundant.
But what exactly is the object in question, the physical structure, the machinery that has made all this social and cultural change possible? What exactly is the Internet, anyway? If we could somehow decant its substance into a glass vessel, we would see it gradually settle into three strata. At bottom rests the global communications network. This substrate predates the Internet itself by roughly a century, built first in copper or iron wire, but since overlaid with coaxial cables, microwave relays, optical fibers, and cellular radio.
The next layer up consists of computers communicating over that global telecommunications system using a set of shared languages, or protocols. Among the most fundamental are the Internet Protocol (IP), Transmission Control Protocol (TCP), and the Border Gateway Protocol (BGP). This is the core of the Internet per se, and its concrete expression is the network of specialized computers called routers, responsible for finding a path for a message to travel from its source computer to its destination.
Finally, at the top, are the various applications which people and machines actually use to work and play over the Internet, many of them with their own specialized languages: web browsers, chat apps, video games, day trading applications, etc. To use the Internet, each application need only embed a message in a format which the routers can understand. That message could be a move in a game of chess, a tiny fragment of a movie, or a request to transfer money from one bank account to another – the routers don’t care and will treat them all just the same.5
Our story will weave these three threads together to tell the story of the Internet. First, the global telecommunications network. Last, the whole panoply of different software applications that allow users of computers to do fun or useful things over the network. Binding them together, the techniques and protocols for getting diverse computers to talk to one another. The creators of those techniques and protocols built on the achievements of the past (the network) and drew on a dimly imagined vision of the future towards which they were groping (the applications to come).
In addition to these creators themselves, the state will be ever present as an actor in our story. Most especially at the level of the telecommunications networks, which were all either themselves government-operated or subject to strict regulatory oversight. Which brings us back to AT&T. Distasteful as they may have found it, Taylor, Roberts and their ARPA colleagues were hopelessly entangled with the incumbent telecommunications operators, the foundational stratum of the future Internet. The functioning of their network depended entirely on such services. How then, to account for their hostility, for their belief that ARPANET represented a new world fundamentally at odds with the backwards-looking incumbents of telecommunications?
In truth, philosophical, not temporal, distance separated the two groups. AT&T executives and engineers saw themselves as the custodians of a vast, complex machine that provided reliable, universal communications services from one person to another. All the equipment in between was the responsibility of the Bell System. The architects of the ARPANET, however, saw that system as a simple conduit for arbitrary bits of data, and believed its operators had no business meddling in questions of how that data was generated and used at each end of the wire.
We must begin, then, with the story of how this philosophical impasse over the nature of American telecommunications was resolved, by the might of the United States government.
- Computer History Museum, “Interview with Lawrence G. ‘Larry’ Roberts,” June 1988. ↩
- Keenan Mayo and Peter Newcomb, “An Oral History of the Internet,” Vanity Fair (July 2008) ↩
- A long session of aimless time-wasting on the Internet seems to leave one’s mind in a similar state to that of one’s body after eating an entire bag of Doritos in one sitting. ↩
- Among other things, those hoards have been hugely important to the neural-net renaissance of the past decade. For example, “ImageNet Classification with Deep Convolutional Neural Networks” (2012), the foundational paper of that renaissance, depended on a database of millions of images harvested from the Internet and labeled by a distributed group of anonymous people recruited over the Internet via Amazon’s Mechanical Turk service. ↩
- To a first approximation, anyway. We may get to the question of net neutrality later, should our story carry us so far. ↩