It would be hard to find anyone in information technology with the track record of Larry Smarr. In 1985, he launched the National Center for Supercomputing Applications, a federally sponsored research facility, where he led the development of what would become today’s Internet backbone. Years later he worked with a then unknown student, Marc Andreessen, to develop and commercialize the technology that would jump-start the Internet revolution: the Mosaic Web browser.
Now Smarr, perched in a seventh-floor office overlooking a eucalyptus grove on the University of California, San Diego, campus, directs the California Institute for Telecommunications and Information Technology. His is one of four state institutes set up last year to foster collaboration among government, industry and academe and spur advances in information technology, biotech, nanotechnology and the Internet-helping ensure California remains a high-tech powerhouse.
Throughout his illustrious career, Smarr, 52, has been a master facilitator, bringing people and institutions together to work on key technological challenges. At this new institute, he could face the greatest test to date of his consensus-building skills. Drawing upon the academic prowess of UC San Diego and UC Irvine, the center links three levels of government-state, federal and local-200 university faculty members, a slew of cutting-edge companies and the community at large to explore how next-generation Internet technologies will transform transportation, medicine and the environment. Freelance writer Eric Pfeifer braved T-shirt temperatures and calm skies to visit Smarr and hear his plans to spend $400 million in four years to create what he calls a massive “living laboratory.”
TR: Okay, so what exactly is a massive “living laboratory?”
SMARR: Well, a number of S-curves will define the development of the Internet over the next five or ten years. Technology goes from a research phase to early adopters and then shoots up, so it seems like everyone is using it, and then it flattens out and becomes mature-a process which looks like a stylized letter S. This institute will be very critical in the first part of this curve-the research phase. We will discover things and fool around with them until they get to the stage where someone says, “Let’s do a startup.” So our goal is to weave together emerging technologies such as the wireless Internet, nanotechnology, chemical sensors and sensor nets, and digitally enabled genomic medicine. We are literally taking technologies and building them in the campus and the community. We will have a living laboratory of what it’s like in the future today, three to five years before it becomes mass market.
TR: What’s the primary benefit of bringing these all under one roof?
SMARR: If you have computer scientists and electrical engineers just studying technology, you may miss the big payoff for California and society-which is what will people do differently when the technology comes into being. How will our quality of life change? How will the economy of California change?
Take wireless technology. For 30 years we’ve been exponentially increasing the number of Internet addresses on the wired Internet. In the next three to five years, we will see more Internet wireless addresses come online than we created in the previous 30 years on the wired Internet. So for example, Hewlett-Packard donated 500 pocket PC devices-Jornadas-which will be equipped with wireless cards, so students will be able to access both local- and wide-area networks. So imagine that you have a pocket PC that you are taking notes on, and it has a wireless card, and I say, “Well, look at this.” I give you a URL, and you pop it up and start working on it and walk out of the office. Well, as long as you are on this floor, you still have local access, but the minute you get in the elevator and head downstairs, your screen goes blank. With the integrated local-area and cellular network we are testing, once you’re at the fringe of the local-area network, you will be able to pick up seamlessly onto the wide-area network.
You’ll also have a map that will come up on your handheld, and when you’re e-mailing your friend on campus, you’ll be able to see literally that your friend is just behind the tree. So imagine crossing instant messaging with geography. With geolocation you have the opportunity for a meeting in physical space or a cybermeeting. Right now, 500 first-year freshmen in computer science at UCSD have these Jornadas, and this will be increased next year. UC Irvine is also beginning experiments with some of their students.
TR: But in a “living laboratory,” you don’t know exactly how these students are going to use the new wireless technology.
SMARR: We don’t want to know. We want to discover that. Let me give you an example. A decade ago, campuses all over the country dug up their quads and put in coaxial cable and then fiber to create what became the broadband networks inside the campus and between the campuses. We all thought this new network would enable supercomputing or virtual telescopes or something. What it enabled was Napster.
I can’t tell you how many national meetings I was in over the last 15 years. Not once did anybody get up and say the reason we’re building broadband networks on the campuses is to enable sharing of MP3s. So that was a discovered application of broadband networks that only could have come about when you had this very broad set of geographically separated students who had a common interest in listening to music. So that’s what we want to do-let the kids figure it out. What’s this good for? What are the new services?
TR: But it’s not just students. Corporations also want some answers. It must be tough trying to balance the needs of corporate America with higher education.
SMARR: Yes, there is a lot of treacherous territory in finding the right balance between industry and academia, which have different sets of values and goals. We are really bringing a new kind of culture to academia. Now, the good news is that the 21st century is very much about this global virtual teaming. But the danger is the barriers will be too great to overcome.
Ultimately, we are trying to enable a crisper technology transfer. A lot of people on campus would love to be more involved with industry and entrepreneurial if they just knew how. And if you don’t have the whole economic ecology present, then students and professors feel like there’s no natural way for them to do this on campus, so there’s more of a tendency for the corporate guys to come in the dark of night and steal some of your best professors and students.
TR: What is government’s role in all this?
SMARR: California governor Gray Davis put up $100 million for each of the four institutes. But he said you have to bring two dollars to the table for every one dollar I’ve given you. He told us that we could get the matching dollars by going after industry, going out to private individuals, or we could go to the federal government. And that’s what we’ve done. Now, the majority of those dollars have been from industry because that’s the shorter-term way to do it. Qualcomm says, here’s an antenna, another company will say, here’s a bunch of wireless devices, and HP will come in and say, here’s Jornadas-which is why we can move very quickly. But at the national level you have to show how all 50 states will benefit. This means you have to go at a slower pace. At the moment, we have proposals which total about $100 million into the federal government, and we expect that to grow over time. These proposals cover a wide range of topics: quantum computing, materials research, sensor development, optical networks and biomedical imaging.
TR: You’ve worked with many emerging technologies. What characteristics do you think a new technology should possess if it is going to be commercially successful?
SMARR: I guess most of them have been standards based. If you go back to the Internet itself, it was TCP/IP based, and the national supercomputer centers were able to use that protocol to link themselves, and then the NSF [National Science Foundation] funded the regional centers and campuses and so forth. It could self-propagate. The Web was the same thing. Standards mean you can decentralize the build-out of a large-scale system; you don’t have to do top-down management.
TR: Which technologies that you’re working on are closest to commercialization?
SMARR: That’s difficult to say. Probably 60 percent of my budget is for things that are going to happen in the next two to five years, for example optical networks or intelligent transportation. Thirty percent of my budget is going to be on stuff that is five to ten years out, which is a lot of the smart sensors. Then 10 percent is for things that are really over the horizon, such as quantum computing and quantum communication.
TR: Let’s start with what’s coming down the line first-intelligent transportation.
SMARR: You’ve got this amazing thing today where a modern car has maybe 20 microprocessors and 60 sensors in it-none of which are connected to the Internet. We’ve got a hundred million vehicles. What if each of those cars had an Internet address? We’d know exactly where each car is and how fast it is going. If we have enough of these as tracer particles, then we have a good idea of the state of traffic as a whole. So you can begin to imagine your car telling you, “Don’t go on the I-5, go up on I-15.” You begin to imagine real-time traffic management.
Will Recker, director of UC Irvine’s Institute of Transportation Studies, has worked with Toyota to get 50 cars. His plan is to have a pool of cars, all zero- or low-emission vehicles, at stations of rail lines, like the Amtrak Pacific Surfliner that stops in Irvine and San Diego. The idea is that a person gets off the train, walks up to the car, swipes a smart card, is authenticated and drives home. We will outfit the cars with GPS and other telematic sensors being designed in collaboration with UCSD. So these cars will be sampling the local traffic flow and wirelessly reporting back while the person is driving.
TR: What other smart sensors are you developing, and how will they ultimately be connected wirelessly to the Internet?
SMARR: If you look at a company like Nanogen, they are able to shrink a DNA array to about two millimeters square. What if you could take that and put it with system-on-a-chip technology so that you could integrate-on silicon-a microsensor along with embedded software, a large memory and an RF [radio frequency] or laser transmitter? And what if you could shrink that down to [two-thirds of a centimeter]? What that amounts to is a super sensor. It is not just detecting accelerations or biological and chemical signatures, but also doing vast amounts of computing.
We’re doing a project with San Diego State University. They manage an ecological reserve in northeast San Diego County where the institute will be testing out these smart sensors. One of our other academic partners is the High-Performance Wireless Research and Education Network, an organization funded by the NSF that is building out a set of 45-megabit-per-second, point-to-point wireless Internet links across southern California using the unlicensed spectrum band. That’s arguably 100 times faster bandwidth than any third-generation wireless cellular system you are going to see for five years. This network currently provides a wireless Internet infrastructure to this ecological reserve and sensors, which are measuring everything from atmospheric turbulence to the temperature of ground leaves to warn of potential brush fires. It also has sensors that are listening to bats to help animal behaviorists understand the bats’ life cycle.
Eventually, at this ecological reserve and even across the country, there will be quality-of-water sensors that will monitor the major estuaries, and they will alert the appropriate officials if there is a pollution problem. That’s one of our major goals: to establish an early-warning system for ecological damage. Gradually, bridges will have seismic sensors on them. Highways will have air quality monitors. We are going to have a society where there are sensors everywhere.
TR: This completely changes the physical nature of the Internet.
SMARR: Right now you could argue that humans are very much in the loop on the Internet. Over time, the percentage of traffic on the Internet that touches a human is going to be a very, very tiny fraction. More and more, it will be traffic between computers, between sensors and databases, between embedded processors like in cars or PDAs [personal digital assistants]. The upside to this is we are going to have a much more automated society; we are going to have a world of intelligent agents that go off and work for us during the day or night. Everyone is going to have dozens of personal servants that are software.
TR: How about genomic medicine. Where does it fit into this mix?
SMARR: We are just now getting going on digitally enabled genomic medicine. One of our professors at UCSD just got a $25 million National Institutes of Health grant to set up what’s called BIRN, the Biomedical Informatics Research Network. This will be a national network, initially for brain imaging, involving Harvard, Duke, Caltech, UCLA and UCSD. Right now, there is no central place where all MRIs are stored. UCSD will be the software and network developer for this national repository. People studying genetics will be able to do 3-D visualization. They will be able to fly their way inside of a particular brain, and then by software, we will link portions of the brain to genetic and proteomic databases. So someone flying through a brain will be able to click on the pituitary gland, and up will come those genes that are expressed in the release of its hormones. The beauty of this is that we aren’t doing brain research, we are building an infrastructure into which knowledge can be poured. If this works, and we think it will, then the NIH wants to extend it to other organs and many other laboratories.
TR: You’ve hit two home runs: developing the Web browser and the Internet backbone. Do you think you’ll be able to hit your third one here at the institute?
SMARR: [Laughs.] In each of those cases I helped to catalyze things. I was a midwife as opposed to a father or mother. The best I can do is get a bunch of smart, innovative people together and let them live in the future. The chances are pretty good they will discover new features or services about the infrastructure that will get widely adopted. What I’m trying to do, what I’ve always tried to do, is accelerate the future getting here.
Geoffrey Hinton tells us why he’s now scared of the tech he helped build
“I have suddenly switched my views on whether these things are going to be more intelligent than us.”
Meet the people who use Notion to plan their whole lives
The workplace tool’s appeal extends far beyond organizing work projects. Many users find it’s just as useful for managing their free time.
Learning to code isn’t enough
Historically, learn-to-code efforts have provided opportunities for the few, but new efforts are aiming to be inclusive.
Deep learning pioneer Geoffrey Hinton has quit Google
Hinton will be speaking at EmTech Digital on Wednesday.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.