The Geopolitics of Asian Cyberspace
December 2006

by Ronald Deibert 

What happens to your request when you click on a link to a Web site or send an email?  For most surfers, the Internet experience begins and ends with what happens on the computer screen in front of them.  However, if you follow that email or Web request as it leaves your computer and passes down the fiber-optic cable to the servers and routers of your local Internet Service Provider (ISP), through the Internet Exchange Points (IXPS), international gateways, and on to the undersea trunk cables of telecommunications companies, you will find a complex and largely hidden infrastructure of filters and choke points. 

HARRY HARRISON

Conventional wisdom had it that the Internet was an unstoppable force for liberalization, with nondemocratic states powerless to control this sprawling, seamless network of networks.  But this vast international “underbelly” of the Internet —almost completely invisible to most Internet users—has become an object of geopolitical contestation among states, and a site where political power is being asserted.

Perhaps nowhere is the geopolitical dynamic playing itself out more forcefully than in the vast region of Asia. Home to one of the world’s cyber-superpowers, China, and dozens of newly emerging markets eager to capitalize on the benefits of new information and communication technology (ICT) while limiting negative side-effects for centralized political authority. Sophisticated ICT companies, many from the West, are following the lead of Asian governments, offering a wide range of appropriate products and services.

This exercise of political power in cyberspace by states in Asia, however, is not going uncontested. A swarming resistance movement of tech-savvy citizens is forming to protect the Web as an unfettered forum of freedom of speech and access to information. With the development of new software tools designed to circumvent censorship, they are taking the battle to the Internet’s inner core.

Perhaps the best window on the dark underbelly of the Internet comes from the research of a project I direct: the OpenNet Initiative (ONI)—a collaboration among the Citizen Lab at the University of Toronto, the Berkman Center for Internet & Society at Harvard Law School, the Cambridge Security Programme, the Oxford Internet Institute, and partner NGOs worldwide. The aim of the ONI is to document empirically patterns of Internet censorship and surveillance worldwide using sophisticated means of technically interrogating the Internet directly.  The ONI’s tests are carried out both remotely from North America and the United Kingdom, and in the field by dozens of local researchers. Our reports over the last several years have documented a disturbing increase in the scale, scope and sophistication of Internet censorship practices worldwide, including in Asia.

When the ONI was formed in 2002, only a handful of countries were known to engage in Internet content filtering, most prominently China, Iran and Saudi Arabia.  Now more than four years later, the ONI is presently testing in more than 40 countries worldwide. China is still the world’s most notorious and sophisticated censoring regime. Its filtering system comprises multiple levels of legal regulation and technical control, the latter implemented primarily at the backbone level using specially configured Cisco routers. The system involves numerous state agencies and thousands of public and private personnel, and a dense web of ever-thickening legal restrictions.

However, China is not alone. Among countries that the ONI has researched in Asia, we have technically confirmed Internet content filtering in Burma, Vietnam, the Maldives, Thailand, South Korea, Singapore, Pakistan and India. Although we have not yet conducted tests in North Korea, it is well known that what little Internet exists in the country is heavily filtered. Likewise, Australia filters Web content through official takedown notices issued to ISPs by the government. In Central Asia, we have also identified extensive Internet censorship practices in Uzbekistan, and intermittent or targeted filtering in Kazakhstan, Kyrgyzstan and Tajikistan.

Mission Creep

Although many countries justify their censorship practices as a way to block access to pornography or other culturally sensitive material, our research has documented a large and growing swath of content beyond pornography that is targeted for filtering. 

In China, Burma, Uzbekistan and Vietnam, for example, ISPs routinely filter access to the Web sites of local opposition parties, dissident and independence movements, and human-rights and religious groups.  Free email, Web-hosting, and blogging services; online encyclopedias (such as the popular Wikipedia, in the case of China); privacy and security tools; and instant messaging clients are also targeted for filtering. Other countries appear to be following this path. In Thailand, for example, what started out as an effort to block pornography has been gradually broadened to include politically-sensitive Web sites as well, particularly since the September 2006 military coup. Pakistan began filtering sites that contain imagery offensive to Islam, and now targets sites related to the Balochistan independence movement as well. The Thailand and Pakistan cases may suggest that once the tools of censorship are put in place, the temptation for authorities to employ them secretly for a wide range of ulterior purposes may be large —particularly in circumstances where there is little civilian oversight or accountability—a phenomenon we refer to as Internet censorship “mission creep.”

The ONI has also documented a more concentrated effort among states to target content in local languages, such as Vietnamese, Mandarin, Arabic and Farsi. For example, our in-country tests in China compared search results for keywords in both English and Chinese and found a much higher rate of inaccessibility for contentious Chinese keywords.  Phrases such as “Chinese Labour Party,” “China Democracy Party,” “Party for Freedom and Democracy in China,” and “Inner Mongolian People’s Party” are less likely to be accessible in Chinese than English. The same variation rates of inaccessibility were also found in Vietnam and Iran. Although determining the motivation for such variation is difficult, one might surmise that political authorities may want to target that web content which hits closest to home while leaving English speaking visitors to the country (e.g., journalists, Western human-rights activists) with the impression that censorship is rare.

The increased sophistication of Internet content filtering practices can be attributed, in part, to the services provided by Western (mostly U.S.-based) software and Internet service firms. Whereas once the best and brightest of Silicon Valley were associated with wiring the world, and opening up access to vast stores of information, today they are just as likely to be known for doing the opposite. Although Microsoft, Cisco, Yahoo!, Skype, and Google have all come under scrutiny for colluding with China’s Internet censorship practices, perhaps the most significant, serious, and yet overlooked contribution to Internet censorship by Western corporations comes from the manufacturers of the filtering software used to block content.

Internet security companies like Fortinet, Secure Computing and Websense create off-the-shelf filtering products that block access to categorized lists of Web sites. While these products are primarily marketed to businesses, they have been readily employed by censoring states like Tunisa (Secure Computing), Iran (Secure Computing), Burma (Fortinet), and Yemen (Websense) to block access to politically sensitive content. 

Just like businesses that do not want their employees to view gambling or sport sites on company time, these governments simply tick off those categories of Web sites they do not want their citizens to access, such as “advocacy groups” or “militancy and extremist groups”—two categories in Websense’s database. The former is defined by Websense as “sites that promote change or reform in public policy, public opinion, social practice, economic activities and relationships,” while the latter is defined as “sites that offer information about or promote or are sponsored by groups advocating antigovernment beliefs or action.” 

Digital Deceit

One troubling trend identified has been the lack of accountability and transparency over Internet-content filtering practices by states that censor. While there is certainly a legitimate debate to be had about the balance between a state’s right to cultural sovereignty and the free flow of information, unfortunately most states do not allow such a debate to take place prior to filtering, and have been shown to be deceitful about the content they block and the filtering practices they employ.

While some authorities yield clearly labeled blockpages to users who request banned content, others are not so transparent. In China, for example, ONI researchers found through forensic analysis that China’s backbone routers are configured such that requests for banned content result in a network timeout error. The routers then send packets to the user’s machine effectively blocking that user’s unique IP address for an indefinite period of time such that any further requests for any web content on the same server results in a network timeout error. Among some ISPs in Uzbekistan, requests for search engines or political opposition Web sites are redirected to search engine Web sites. In Tunisia, “spoofed” Web sites are returned instead of blockpages to give the appearance of network errors instead of deliberate filtering. In Kyrgyzstan and Belarus, ONI researchers found troubling patterns of inaccessibility to opposition Web sites during election periods in each of those countries—inaccessibility we traced to denial of service attacks or what appears to be deliberate tampering with domain name servers. The latter cases suggest a new form of “just-in-time” filtering which is less easy to verify or attribute responsibility but just as effective in blocking access to politically sensitive sites. The authorities leave the Internet open, in other words, but turn the tap off only at critical periods, such as during elections or public demonstrations.

Adding to problems concerning lack of transparency and accountability is the use of proprietary filtering software outlined above. The manufacturers that make the commercial products used to block access to information treat their blacklists as trade secrets, revealing only general categories or sample Web sites to the public. This combination of proprietary software tools and nondemocratic regimes creates a kind of double bind of unaccountability.

Protecting the Net

Perhaps the most powerful way to protect the net, however, is by taking the battle to the underbelly of the Internet itself —by creating technologies that help secure, rather than undermine—access to information and freedom of speech.

At the Citizen Lab at the University of Toronto, our contribution comes in the form of a simple censorship circumvention program called psiphon, released on Dec. 1, 2006. The software works by having people who live in uncensored countries, like Canada or the U.S., install a psiphon server on their home computer, and then give the connection information to a few friends or family members that live inside censored countries. Because each psiphon node runs separately on private computers, and the connections among psiphon users and servers are encrypted, it is extremely difficult for authorities to block.

Psiphon will allow individuals to bypass their government’s filters, and connect to the Internet the way it was originally meant to be experienced. Rather than set the bar at the lowest common country denominator, psiphon sets it at the highest, most open Internet standard, allowing individuals themselves to determine what information they can or will not access, rather than deferring such a choice to bureaucratic elites. Yet psiphon and other similar software are not magic bullets that will end Internet censorship once and for all. The job of preserving the Internet as an open global commons of information is a constant cat and mouse game between citizens who want access to information and authorities that want to control and limit it. 

But even so, these grass-roots technological efforts—that shore up the Internet’s infrastructure as a medium that supports, rather than detracts from, human rights—may be citizens’ best weapons in the ongoing geopolitical battles over cyberspace.

Mr. Deibert is director of the Citizen Lab at the Munk Centre for International Studies at the University of Toronto, and a principal investigator of the OpenNet Initiative. The psiphon software can be downloaded from http://psiphon.civisec.org/.

back to top
 

Dow Jones Links

• WSJ.com
• WSJA.com
• Chinese WSJ.com
• Dow Jones