The next "Big One" has finally hit the San Francisco Bay Area. A massive earthquake has left entire city blocks in ruins. There is no water or electricity. Disaster recovery teams are roaming the city, but coordination is nearly impossible, as many of the cell phone towers have toppled.
Halfway across the globe, a group of soldiers are separated from their platoon in a crumbling building and are surrounded by enemy forces. They attempt to radio their commanding officer for backup only to hear static, and the war-zone equivalent of a dropped call.
In both of these instances, the answer to the quintessential question of the wireless age--"Can you hear me now?"--is a dangerous and terrifying "no."
UC Santa Cruz engineering professor J. J. Garcia-Luna-Aceves is hoping to change that through a multidisciplinary look at the science, technology, and even social side of networking. If he and his colleagues succeed, the benefits would quickly trickle into our everyday lives, not only eliminating cellular "dead zones" in our cities but helping create a world where mobile video chat becomes commonplace, our myriad devices link up seamlessly, and smart objects communicate with each other, and with us.
In this vision, cyberspace shifts from a place you go to through your personal computer, to an overlay on top of our existing reality where the information you need is proactively delivered right to you, wherever you are.
Rethinking networking
 |
|
J. J. Garcia-Luna-Aceves |
"We've been designing wireless networks the same way we design wired networks," Garcia-Luna-Aceves says. "As a result, they don't work as well as they could. So we're rethinking networking from the ground up."
While wireless networks, from cell phone services to WiFi, have become nearly ubiquitous in the last decade, their underlying network technology dates back to the 1960s. It was then that the United States Department of Defense funded development of the ARPANET (Advanced Research Projects Agency Network), the predecessor to today's Internet.
For the ARPANET, and the Internet, to work effectively, the network had to permit any node (or person) to easily share information with any other node. All information is treated the same way, and transmitted with the same likelihood of success. Indeed, the ARPANET's architects developed protocols and systems that provided for just that, and the Internet of today still runs on their innovations. The rub is that while those protocols are well-suited for wired networks, they fall short when you clip the cables.
For starters, wireless networks have lower bandwidth than the fiber-optic lines that are the main arteries of the Internet. Meanwhile, the layouts, or topology, of wireless networks are constantly changing as people and devices move between cell towers, for example, or the radio channels are sullied by interference, making them less reliable.
The social factor
According to Garcia-Luna-Aceves, the first step in addressing these issues is to strip away all the technology and reconsider what is meant by a network in the first place. Historically, researchers have focused solely on improving communications networks by developing new protocols and methods for routing data around uniformly, just as the creators of the Internet did.
Indeed, Garcia-Luna-Aceves and UCSC engineering colleagues Hamid Sadjadpour and Katia Obraczka recently led a multi-university project called Dynamic Ad-hoc Wireless Networks (DAWN) to develop and test new routing protocols to increase the capacities of wireless networks.
For example, the network could be smart enough to pre-fetch information that a user may want and then hold it until it's needed, using the network more efficiently. For the user, the data would be right there instead of having to point, click, and wait. Of course, determining what to store, and for how long, is no small feat for a computer algorithm.
And no matter how smart the software, a network, Garcia-Luna-Aceves explains, is much more than the digital technology behind the curtain. A network is also the people who use it, and the information that's exchanged across it. Those social factors and information demands should be taken into account when building wireless networks where almost every resource--from bandwidth to battery life--is at a premium.
"The notion that we need to enable everyone to talk one-to-one with everyone else, and maintain the routes to all these sites and nodes that we never use, is a big problem," Garcia-Luna-Aceves says. "So now we are studying how the flows in a network result from common interests and needs."
This idea is somewhat akin to the concept that you don't build highways to everywhere--less-traveled-to destinations are best served by small roads.
Currently, the researchers are exploring how to map a social network on top of a network infrastructure. For example, in a military setting, there's a chain of command, which is a form of social network.
So the network needn't require that every device be able to talk to every other device. Meanwhile, many military communication systems require a great deal of bandwidth, which also must be accounted for in the network architecture.
"Only by figuring out the social network and information network overlays can we start talking about increasing the capacity of the network and delivering quality information to the users," Garcia-Luna-Aceves says. "One of the challenges is to develop a mathematical theory of networking that encompasses the whole range of networks, including wired networks, in a holistic way."
Going ad-hoc
The Network Sciences Institute at UCSC is funded by the U.S. Army Research Laboratory through a new 10-year, $35.5 million grant for the Communications Networks Academic Research Center, a collaboration involving Pennsylvania State University, UCSC, UC Davis, and other organizations. The army is particularly interested in the mobile ad-hoc networks that can be deployed in places where there is no mobile infrastructure.
In these systems, there are no central nodes in the network. Rather, each device passes data from one to another bucket-brigade style until the information reaches its final destination. Of course, a similar approach could be used to establish robust communications in disaster areas or even regions where wireless infrastructure was never built.
"J. J. has always found a way to address critical military needs, such that the U.S.'s national security can be increased, but his work has also found relevance in commercial environments," says James Freebersyser, director of advanced systems development for Raytheon BBN Technologies, based in Cambridge, Mass. "From my perspective, his ability to operate in both spheres is unique."
These kinds of ad-hoc wireless networks are also a key foundation for next-generation sensor networks, consisting of tiny wireless transceivers outfitted with sensors for myriad applications such as environmental monitoring, diagnosing a building's structural integrity, or reducing energy consumption in the home. For example, a wireless smart thermostat in your house might link with temperature, motion, and sunlight monitors in every room to automatically create microclimates that keep you comfortable while also lowering your energy bill.
Outside of your house, pervasive networks could provide access to a variety of location-enabled services, where restaurant reviews or local news is delivered in context through "augmented reality" applications that are much more elegant and accurate than today's early offerings. Your cell phone, tablet computer, and digital camera could link automatically to a wall-size display or office printer as you pass by, or instantly interface with the entertainment system in your car. This is the transformation in networking that Garcia-Luna-Aceves is hoping to spur.
"The network and technology should disappear so that it operates seamlessly in the background to deliver services the users actually want," Garcia-Luna-Aceves says. "Only then will we have a revolution in networking."
David Pescovitz is co-editor of the popular blog BoingBoing.net and a research director at Institute for the Future, a nonprofit forecasting think tank in Palo Alto.
