Helping dissenters evade foreign eavesdropping

In 1998, Yahel Ben-David made his first trip from Israel to India's Kangra Valley, a lush stretch of fields and forests in the shadow of the lower Himalayas.

He traveled to Dharamsala, the stronghold of the Tibetan government in exile and home of the Dalai Lama, to help them set up online security systems. This would increasingly include consulting on methods to allow independence movement organizers to communicate anonymously, avoiding eavesdropping by a Chinese government that has regularly used violence to quash dissent in the territory.

Fascinated by the people and cause, Ben-David ended up spending longer periods in India, until moving there around 2004. But he remained conflicted about his advisory role, never convinced there was a solution that could be relied upon when lives were at stake.

"Preserving anonymity is the No. 1 priority for dissent activity," he said. "Otherwise you get shot."

Ben-David is now a doctoral student in computer science at UC Berkeley, where he is working with a group of researchers who hope to improve the options and odds for activists.

They're developing an Android app designed to allow anonymous communications through a Twitter-like microblogging service, even when oppressive governments shut down Internet and cellular infrastructure. They call the technology Rangzen, the Tibetan word for freedom or liberty. They believe the service could work across hundreds of thousands of devices.

The need for such alternative communications services became especially obvious during the Arab Spring. Nations like Egypt, Syria and Libya blocked websites, shut down Internet service providers, cut off phone networks, cracked into Facebook pages and more.

Ben-David also had some firsthand experience with these matters through earlier work in Israeli intelligence. He declines to discuss the period in detail, but says it provided "some perspective on how governments use and abuse their strengths when it comes to control of infrastructure and communications."

Last month, he and the team of Cal computer scientists and engineers published a report explaining their core concept.

The starting point is what is called a mobile-mesh network, which takes advantage of radio signals in cell phones like Bluetooth, Wi-Fi or NFC to connect to other phones, independent of a cellular signal or Internet connection. Such networks can facilitate the exchange of text messages, voice calls, images and even short video clips.

This technique isn't novel in itself. Other researchers have advanced it as a tool for people in need of impromptu communications networks in the wake of natural disasters. Young men and women in Saudi Arabia reportedly use Bluetooth to flirt across strictly segregated public spaces, like restaurants.

But these networks are susceptible to attacks, from agents either spreading misinformation or flooding the network with messages in a denial-of-service attack.

Sorting messages

The key technical advance in the group's research is a prioritization algorithm that ranks which messages get through and how quickly, based on levels of trust. Trust is a tricky thing when every participant is anonymous, but the basic approach is similar to one used in social networks and online comment ranking: The more people who connect with you directly, and the more connections you have in common with others, the more trusted you are within the network. In turn, your messages are prioritized in the system.

"It's actually hard for someone who isn't trustworthy to build up trust with lots of people who are trustworthy," said Eric Brewer, a co-author of the paper and computer science professor at UC Berkeley. Brewer is also Ben-David's doctoral adviser and a Google executive who works on improving Internet connectivity in the developing world.

The Internet giant is also cited in the paper. The work was enabled through a grant from the U.S. State Department's Bureau of Democracy, Human Rights and Labor, as part of its Internet Freedom initiative.

The other authors include Scott Shenker, Sebastian Benthall and Giulia Fanti, whom Ben-David credits with the key technical insight in the paper. The group is looking to hire additional researchers.

No one says this is a perfect solution. In fact, Brewer himself is quick to qualify their accomplishment, calling it "an interesting step in a very tricky problem."

In some regions, smartphones aren't yet ubiquitous enough for this to work on a large scale. It's possible the mere proximity required could reveal people's identities. And it could remain difficult to get critical messages across national borders.

Major concerns

Some have graver concerns, including Gustaf Björksten, technology director with Access Now, a human rights organization focused on improving access to communications technology. He believes the approach itself is flawed because mobile devices are inherently identifiable.

"All devices leak information in order to participate on the network, regardless of which radio is being used in the mobile device," he wrote in an e-mail. "SIM cards have unique IDs, the handsets have unique IDs, network interfaces have unique MAC addresses, etc. These can all be tied to the user, and given that in many countries to buy a cellular phone, and/or a SIM card, you have to give full details of your identity ... this makes identifying users trivial for governments in collusion with the carriers."

But others, including some with on-the-ground experience in dangerous activism work, see considerable promise in these dissent networks.

Ahmed Salah played a key role in the Egyptian revolution of 2011, and has been arrested and tortured on several occasions. As smartphones become more common in Egypt, a tool like the one the UC team is developing could allow organizers to coordinate over greater distances, potentially swelling participation in uprisings, he said.

Without safe ways of communicating, he and other organizers were forced to travel long distances to meet people in person and slowly spread the word about upcoming plans.

"We had to do things the hard way," he said. "Through personal contact, travel and training. You don't know the person you're going to meet. You could be meeting an agent. It leaves the door open to lots of risks."

James Temple is a San Francisco Chronicle columnist. Dot-Commentary appears three days a week. E-mail: jtemple@sfchronicle.com Twitter: @jtemple