After founding search engine pioneer Inktomi Corp., Eric Brewer became a paper billionaire — before the market crash. That got him thinking about how he could effect change in the world. Today, as associate professor of computer science at the University of California at Berkeley, Brewer is involved in the ICT4B (Information and Communications Technology for Billions) project, which seeks to engineer high-tech solutions for the Third World so as to create sustainable businesses not dependent on charity. In a small office across from the Berkeley campus, Brewer spoke with passion and zeal about his work.

EE Times: What exactly is ICT4B?

Eric Brewer: It's an NSF-funded project that combines engineering and social science. Tier [Technology and Infrastructure for Emerging Regions] is a part of the project that I manage on a day-to-day basis. ICT4B has six faculty, including two in social sciences, and on the order of six students.

EET: Why are you working on technology for the Third World?

Brewer: I started thinking about it seriously at the World Economic Forum in Davos, Switzerland, in 2000. There were some very high-quality people from developing regions in attendance, and they were able to explain what their needs are. It became obvious that First World technology is not a good fit for those needs, for a lot of reasons — the big three are power, money and complexity.

There have been projects that bring technology to developing regions, but they've had three problems. First, they have to use existing, off-the-shelf technology, which may not be a good fit. Second, they tend to be based on donations, which means projects aren't sustainable. Third, there tends to be no infrastructure in these countries, and for a single project you can't afford to provide one. [But] all three problems are addressable.

EET: What's the basic idea behind ICT4B?

Brewer: There is a three-part hypothesis: that there is a role technology can play in developing regions, that First World technology is a bad fit and that we need a better research agenda. Universities can close the gap between the technologies that exist and the technologies that are needed.

EET: You aim to create sustainable projects not based on charity. Is that possible, given that people in developing regions have low incomes?

Brewer: One answer is that there are a lot of them. The aggregate wealth among the poor is not trivial. People in developing regions actually do have disposable income; in fact, they pay a lot for the things they buy. Another answer is that we don't have to think about personal devices. Farmers can't afford tractors, but villages can. One example is Grameen Telecom, which has reached about 52,000 villages in Bangladesh with cell phones. A woman takes out a loan to buy a cell phone and she pays it back by renting minutes to neighbors. Her income tends to be three to 10 times the average village income. That is a very profitable business that grew very quickly.

EET: What kinds of problems can ICT4B help solve?

Brewer: Almost every problem we've looked at has some potential for technology. In Tanzania, the child mortality rate was cut in half just by using computers to track why kids are dying. In West Africa, using a sensor network and satellites, they were able to figure out when to spray larva sites for black fly. By spraying at the right time and the right place, they freed up 100,000 square miles that were previously unfarmable — enough to feed 17 million people. That was a World Bank project.

EET: What kinds of engineering challenges need to be solved first?

Brewer: The first is rural connectivity. In low-density areas, there is no viable network solution at the moment. We're looking at a variety of technologies. The one we're working on the hardest is a modified Wi-Fi. It's a nice choice because it's very inexpensive. The problem is that the hardware was designed to work for a few hundred feet at most. The big issue is how you cover 20 kilometers. We're working on that through the MAC [media-access controller] layer, because we don't want to change the hardware.

EET: Is this the long-distance MAC project described at the Tier Web site (http://tier.cs.berkeley.edu)?

Brewer: Yes. We're working on better utilization of the channel for long distance. The normal protocol has a lot of round trips. We had to get rid of the exponential backoff, which leaves the channel idle for long periods of time if you go a long ways. The other thing we're working on is putting multiple antennas on the same pole together. We're doing all our work in software. Hardware is kind of off-limits, because we'll never get to the $20 price points of existing cards if we do our own cards. I think once we've done the software, we'll have some recommendations on how to change the hardware. We've had some discussions with Intel and Atheros.

EET: Are there other challenges?

Brewer: The next topic is intermittent networking. IP [Internet Protocol] assumes that all routers are connected all the time. But in developing countries, due to power and other issues, links are up and down all the time. What's needed is more of a "store and forward" style network. It's like e-mail, where it doesn't matter if the network is down for a while, because it knows to send the e-mail when the system is up. We're generalizing that idea and getting applications to work on an unreliable network. It's fine for education, weather, health care — and it may be 10 times cheaper than a full Internet.

We're also working on user interfaces for speech recognition for unusual languages. Right now we're working on Tamil, spoken in Sri Lanka and in Tamil Nadu state in India. It's too hard to do general-purpose recognition, but we can learn 40 or 100 words that will allow someone to talk to a computer. It's for low-literacy interaction, and also because speech is a lower-cost interface than a keyboard and screen.

EET: What projects are under way in India?

Brewer: We have three big ones at the moment. In Kerala state, we're setting up an e-government wireless network. We're trying to understand how kiosks can deliver services like land records and payments. We have about 300 already set up.

In Tamil Nadu, we're working with the M.S. Swaminathan Foundation. We're doing speech recognition there in kiosks, and it's the first place we've deployed some wireless links [for] things like weather reports for fishermen. In fact, the system was used to broadcast a warning after the first wave of the [Dec. 26, 2004] tsunami hit. That village had only three deaths because people were off the beach.

The third project is with the Aravind Eye Hospital in Tamil Nadu. They can do profitable cataract surgery there for $10, but they can't always get people to come to the hospital. We want to deploy enough rural kiosks so everyone can be within 20 kilometers of a rural health center. In our proposal, there will be something like 10 kiosks connected over Wi-Fi links. In the kiosk, there will be a practitioner who can do basic stuff and a doctor on consultation by videoconferencing.

EET: How often do you go to India?

Brewer: I go every six months, in January and July. Most of my work is developing relationships with NGOs [nongovernmental organizations] or government. Usually I spend all my time in meetings, but occasionally I get to do some fun stuff.

EET: What kind of interest have you received from UC students?

Brewer: Incredible. This is the first project for which I've had volunteer students as well as volunteers from the community. There are more people working on the project than I can fund.

EET: Why are you personally involved in this effort?

Brewer: There was a period when Inktomi was worth $25 billion, and I was the largest single shareholder. I was a billionaire on paper, and when you have a billion dollars, you start to think about what kinds of things you can change. There aren't many places where you can make a real impact like developing regions.

I don't have a billion dollars anymore, but there are other ways to contribute. I don't need the money from this job, but I work it because I love the research. If I can't affect at least 100 million people it's not worth my time.

It's an open question as to whether we'll make a difference. But if you look at the First World over the past 100 years, significant changes in the quality of life have primarily come from technology.