The World Wide Web isn't as worldwide as it used to be, according to research presented earlier today by the Open Net Initiative (ONI).

Many governments are quietly preventing their citizens from viewing specific types of content -- including material that these governments find objectionable for political, social, or military reasons, according to the ONI.

The ONI -- a partnership among groups at Cambridge, Oxford, Harvard, and Toronto universities, has been studying the use of content filtering technology and the availability of information in 41 countries for about a year. Its full report will not be available until the fall, but the group shared some of its initial findings with the media this morning.

The ONI is attempting to collect hard data on the censorship of Internet information, a practice that has frequently been discussed anecdotally but never measured empirically.

"It's a practice that has been difficult for the citizens of those countries to fight against, because often, they don't know that it's being done," says Rafal Rohonzinski, director of the advanced research group in Cambridge's Security Program. Many governments disguise their site blocks as network errors or practice filtering at the search level, so that end users never know their data is being censored, he says.

In the study, 25 of 41 countries were found to be practicing some sort of content filtering or site blocking, often for political reasons. While some countries simply disallow traffic from specific IP addresses, others use sophisticated content filtering tools, such as Websense or Fortinet, to selectively restrict access to data from all over the Web, Rohozinski says.

The ONI wants to raise consciousness about Internet censorship in hopes that citizens and governments will do something to regulate it. "Right now, there are no international laws that deal with the individual's right to information," Rohozinski observes. "Human rights groups can't act, because these governments aren't breaking any laws."

The study focuses on 41 countries where not much was previously known about filtering practices. "Burma, China, Iran, Syria, Tunisia, and Vietnam were found to engage in politically motivated filtering," it says. "Iran, Saudi Arabia, and Tunisia were found to practice substantial social content filtering. Burma, China, Iran, Pakistan, and South Korea were found to filter Websites associated with extremism and separatism for national security reasons."

"No evidence of filtering was found in 14 countries, including Afghanistan, Egypt, Iraq, Israel, West Bank and Gaza, Malaysia, Nepal, Venezuela, and Zimbabwe, many of which one might expect to find Internet filtering," the study says.

The study did not investigate countries where content filtering is already known to be widely used by groups and corporations, such as the United States, Canada, and much of Western Europe. "I can't say for sure, because we didn't study them, but my feeling is that it's unlikely the governments in those countries are doing filtering," Rohozinski says.

The study is currently using network-based detection systems to discover the presence of content filtering technology, but some governments will soon become too sophisticated to be caught that way, Rohozinski says. Some governments already are moving to search-based filtering, which is much harder to detect, while others are simply hiring computer criminals to selectively launch denial of service attacks on sites they find objectionable.

Some observers say the Russian government may have been responsible for recent DOS attacks on government servers in Estonia, for example. (See DOS Gets Political in Estonia.) The ONI has witnessed similar denial of service attacks during elections in eastern European countries, Rohozinski says.

A less obvious tactic is to do content filtering in the country's most popular points of access, such as workplaces and cybercafes, Rohozinski says. "There are lots of countries where most people don't have a home computer, so they get most of their information from work or public places."

The ONI soon will begin a project designed to track governments' use of surveillance technology to monitor Web behavior. "In some cases, that is a way to do selective prosecution, without doing nationwide censorship of data," Rohozinski says.

— Tim Wilson, Site Editor, Dark Reading