Tor

In the 1990s, the U.S. Naval Research Laboratory was working on a problem that seemed quite removed from public concerns: how to protect the communications of its agents around the world? The solution they developed was called Onion Routing. No one imagined then that this military technology would become one of the most important tools for defending privacy on the Internet.

The technical principle is ingenious. Each message passes through several intermediate servers, encrypted in successive layers like the peels of an onion. At each step, a server removes one layer and forwards the packet to the next relay, never knowing either the origin or the final destination. An observer intercepting the traffic sees only one node among others, unable to trace the complete chain.

This era also saw the emergence of cypherpunks, those libertarian technologists convinced that encryption represented a bulwark against state surveillance. The meeting between these activists and military researchers at a university seminar created an unlikely alliance. The military needed a vast network of civilian users to drown their sensitive communications in abundant traffic. The cypherpunks wanted to democratize privacy protection technologies. Each party found what it needed.

In 2002, the protocol was reimplemented for the general public and released under a free license. This source code transparency mattered greatly: how could one trust a security tool funded by the U.S. State Department without being able to verify its operation? The community could now audit it freely.

The early days raised thorny technical questions. Should they add dummy data that masks genuine communications? Theoretical security would benefit, but connections would become unbearably slow. The developers made a choice: a network that’s too slow would never attract enough users. Yet Tor’s strength relies precisely on numbers: the more diverse the profiles, the better sensitive communications blend into the mass of daily traffic.

This philosophy guided the entire evolution of the project. Tor had to remain accessible to the greatest number while maintaining a high level of security. The gamble paid off. Journalists protecting their sources, human rights activists in authoritarian regimes, companies concerned about their trade secrets, ordinary citizens refusing advertising tracking: all found in this network an answer to their needs.

The network also makes it possible to host hidden services, recognizable by their ".onion" addresses. The physical location of these sites is masked, which raises controversies. Silk Road, that illegal marketplace shut down by the FBI, used Tor. But reducing the network to its criminal uses would be dishonest. Edward Snowden’s revelations in 2013 showed that the NSA was failing to compromise Tor on a large scale, confirming its technical robustness against the world’s most powerful adversary.

The architecture relies on thousands of volunteer servers scattered around the world. Individuals, associations, sometimes universities, run these relays out of conviction: preserving a free Internet. This decentralized infrastructure constitutes a technical bulwark against mass surveillance. It also illustrates that an alternative governance of networks is possible, outside commercial or state logic.

Tor’s major innovation concerns not so much the encryption of content as the protection of metadata. Who talks to whom, when, how many times? This information often reveals more than the content of the messages themselves. By tackling this structural problem, the network has inspired many other privacy protection projects.

Tor’s sustainability rests on a fruitful paradox. The U.S. State Department continues to fund its development to promote freedom of expression in countries that censor the Internet. Activists sometimes use it to escape Western surveillance. This duality, far from weakening the project, strengthens its legitimacy and resilience. Serving state and citizen interests simultaneously is not a contradiction but a condition for survival.

Faced with the limitations of abstract mathematical models, the developers favored empirical observation of real threats and concrete uses. This pragmatic approach has produced an imperfect but effective system against mass surveillance. The initial technical challenges forged this method: better a usable tool than a theoretically perfect but practically unusable solution.

The project shows that an alternative infrastructure can emerge from the margins of the Internet to become a pillar of online freedom of expression. Technical communities can create tools serving the public interest, outside the usual circuits. Onion routing born in a military laboratory has become a digital commons.