Tor can solve both problems; the same proxies that provide anonymous cover for people posting content also become portals for banned websites. When it officially launched five years ago, the Tor network consisted of 30 proxies on two continents; now it has 1,500 on five continents, and hundreds of thousands of active users. And its developers are trying to expand its reach, both abroad and in the United States, because digital barriers and privacy threats affect even the free world. In the United States, for example, libraries and employers often block content, and people’s Web habits can be–and are–recorded for marketing purposes by Internet service providers (ISPs) and by the sites themselves. “The Internet is being carved up and filtered and surveilled,” says Deibert. “The environment is being degraded. So it’s up to citizens to build technologies to [counter these trends]. And that is where I see tools like Tor coming into play. It preserves the Internet as a forum for free information.”
The product of a small nonprofit organization with eight paid developers and a few dozen volunteer security professionals around the world, Tor takes advantage of the fact that Internet traffic consists of two-part packets. The first part contains data–pieces of a Web page you are viewing, or of the photo file or e-mail you are sending. The other consists of the Internet protocol (IP) address of the sending and receiving computer (plus other data, such as the size of the file). Tor uses the latter portion–the addressing information–to build a circuit of encrypted connections through relays on the network (see “Dodging Spies, Data Miners, and Censors” next page). The requisite relays (which collectively serve as proxies) are operated on a volunteer basis at universities such as Boston University and a few corporations, and by computer-security professionals and free-speech advocates around the world. (Many Tor users also use existing technologies, such as HTTPS–a protocol for encrypting and decrypting a user’s page requests and the pages that are returned–to protect the content they are sending and receiving.)
Tor, like the Internet itself, emerged from military research–in this case at the U.S. Naval Research Laboratory in Washington, which built a prototype in the mid-1990s. The military interest was clear: without a way to make Internet traffic anonymous, an agent’s cover could be compromised the minute he or she visited .mil domains using the Internet connection of, say, a hotel. Even if the data were encrypted, anyone watching traffic over the hotel network could quickly figure out that the guest might be associated with the U.S. military. And the problem is hardly limited to hotel networks; IP addresses can be linked to physical locations by a variety of means (ISPs correlate such data with phone numbers, data miners can piece together clues from Internet traffic, and someone outside your house can confirm that you are the source of specific kinds of Internet traffic by “sniffing” data traveling over Wi-Fi). As a Tor presentation puts it, chillingly, what might an insurgent group pay to get a list of Baghdad IP addresses that get e-mail from a .gov or .mil account?
The navy project never emerged from the lab, but it attracted the interest of Roger Dingledine, a cryptographer concerned about a different aspect of Internet privacy: the way ISPs and websites amass databases on people’s browsing and search history. In 2000, at a conference where he was presenting his MIT master’s thesis on anonymous distributed data storage, he met a Naval Research Lab mathematician, Paul Syverson. The two men saw that tools for protecting military agents and tools for protecting Web surfers’ privacy could be one and the same, and together they revived the project with funding from the Defense Advanced Research Projects Agency (DARPA) and the navy.