Jacobs School Researchers Work with HPWREN on Allocating Resources in Wireless Sensor Networks
San Diego, CA, September 12, 2005 -- An interdisciplinary team of researchers from UCSD's Jacobs School of Engineering is teaming with developers of the High Performance Wireless Research and Education Network (HPWREN) to build better tools for managing resources in heterogeneous wireless sensor networks.
"The goal of this collaboration is to design scheduling and routing algorithms capable of supporting Quality of Service [QoS] requirements for users of HPWREN," says Tajana Simunic Rosing, a professor of Computer Science and Engineering (CSE) on the UCSD team. "There are competing demands on the network from different applications that have varying QoS needs. We want to implement tools and strategies to ensure that high-priority traffic can obtain the resources needed to traverse the network, while still being considerate of the needs of lower-priority data."
Simunic Rosing is an expert in resource management of embedded wireless systems. She is collaborating on the project with Electrical and Computer Engineering (ECE) professor Tara Javidi, who has a strong background in stochastic resource allocation in wireless networks.
"Time and again our theoretical research has shown how the stochastic nature of the wireless channel can significantly change the rules-of-thumb of networking," says Javidi. "In this collaborative research, my aim is to take advantage of HPWREN's cyberinfrastructure to put these results into practice."
According to HPWREN principal investigator Hans-Werner Braun, the research activity with Jacobs School faculty and graduate students will be critical to the success of the project, which is funded through an Information Technology Research grant from the National Science Foundation (NSF). "The disparity of resource requirements of the various HPWREN applications is creating an environment where QoS research results can really make a difference," says Braun, a UCSD research scientist affiliated with the San Diego Supercomputer Center. "This is particularly true in rural and remote wireless environments, where 'throwing bandwidth at the problem' is often not a realistic option."
Other user applications have stringent QoS requirements. Timely delivery of real-time seismic data can be crucial to response efforts in the wake of an earthquake. The same is true for timely information on the spread of wildfires, or changes in environmental conditions such as wind gusts -- and rapid response of emergency personnel may depend on early notification of a problem. "In these situations the individual sensor readings may not consume much bandwidth," says Simunic Rosing, noting that a single seismic sensor rarely exceeds 10 kilobits per second. "But the timely delivery of the data, in the midst of other data already present on the network, can be of critical importance."
"We would like to dispel the myth that QoS provisioning is only picking winners and losers in a zero sum game," adds ECE professor Javidi. "Instead, we hope to show that when carefully designed, QoS provisioning even benefits 'low-priority' streams and users as it improves overall performance."
Three graduate students are also involved on the Jacobs School end of the project. ECE Ph.D. candidate Jaewook Shim is focusing on statistical analysis of data traffic on HPWREN, and development of theoretical bounds on the bandwidth improvements that are possible when scheduling data delivery on a wireless channel. A Ph.D. student in CSE, Donghwan Jeong, is currently setting up a small sensor network in Simunic Rosing's lab for testing and measurement of scheduling and routing algorithms, and has started working on policy-based routing. Rounding out the team is Daeseob Lim, an M.S. student in CSE who is developing a simulation environment which he will use to evaluate scheduling and routing algorithms developed by the team.
The NSF-funded network includes backbone nodes at UCSD and San Diego State University, and a number of hard-to-reach areas in remote environments. HPWREN is used for network analysis research, with an additional strong emphasis on networking requirements for first responders in remote areas. It also created a high-speed wireless Internet collaboration with field researchers from various disciplines (such as astronomy, ecology, and geophysics), as well as for educational opportunities related to rural Native American learning centers.