Every day that computer sciences professor Lili Qiu works to find new ways to make wireless connections faster, more reliable and more accessible is another day that the challenges, and the possibilities, of the wireless world expand.
Since arriving at the university from Microsoft in 2005, Qiu has completed a general wireless interference model that should make it possible to systematically optimize performance for real wireless networks. She’s discovered new security holes that “greedy” users might employ to trick wireless networks into giving them more bandwidth than they deserve—crowding out honest users—and she’s found ways to patch those holes. She’s helped build virtual tunnels across networks and service providers in order to facilitate applications—like Internet-based phone service, video conferencing, and online trading—that depend on uncorrupted and uninterrupted connections. And she’s begun looking at the kinds of challenges that arise as more people play multi-user video games from wireless connections.
Qiu is also developing software to more efficiently transmit information through “multi-hop wireless mesh networks,” which are set up to pass information along through numerous wireless nodes (usually computers) before reaching an access point to the wired infrastructure of phone and cable lines.
Such mesh networks, which were initially developed so that the military could maintain communications in isolated, unwired locales, have been proving increasingly attractive to places—like rural communities, high-density urban downtowns, and developing countries—where it’s too costly or too complicated to set up a fully integrated wired infrastructure.
The software developed by Qiu and her students, nicknamed “SOAR,” has already shown that it can improve efficiency by taking advantage of the fact that a wireless signal is broadcast in all directions, and thus may have a number of other wireless devices within range that are capable of receiving and then re-transmitting a signal. SOAR provides a mesh network with a set of software protocols that allow the network to choose, at each step, the most optimal node in range to pass on a packet of information.
Qiu has been testing her software on a network of 25 wirelessly linked computers in and around her office in the ACES building. Next, she hopes to take advantage of the city of Austin’s multi-hop mesh network for field-testing.
“My work at UT has been about finding a more systematic approach to these
kinds of problems,” says Qiu. “Right now most of the solutions are either
ad hoc—created for a particular network, or they’re so abstract that they’re of questionable utility when dealing with a real system. Ultimately, I plan to develop a complete toolkit that is both solid theoretically and practically effective, and apply it to managing wireless networks at home, on campus, in enterprises, and at medical clinics.”
Comments