Office: 550 Paul G. Allen Center (detailed mail and street addresses).
Phone: 206-616-1069
Fax: 206-543-2969
email: magda at cs...
I am a member of the database group at UW and the co-founder of the Northwest Database Society (NWDS).
I am co-organizing the International Workshop on RFID data management (RFDM'08) to be held in confunction with ICDE 2008 on April 7th 2008 in Cancun, Mexico.
My interests are broadly in the fields of databases and systems. My recent work has been on stream processing systems and federated distributed systems but I like to investigate any problem involving data management: from pervasive and personal data to the mounds of information available on the Web.
We are deploying and experimenting with a building-wide RFID-based tracking infrastructure. The goal of this project is to overcome the intrinsic limitations of the technology to provide useful applications while respecting users' privacy.
Mobile and pervasive applications rely on devices such as RFID antennas or sensors (light, temperature, motion) to provide them information about the physical world. These devices, however, are unreliable. They produce streams of information where portions of data may be missing, duplicated, or erroneous. In this project, we are investigating techniques for detecting and correcting such input data errors. Our goal is to improve the quality of information produced by applications operating on unreliable sensor data. Because frequently errors cannot be corrected with certainty we are investigating probabilistic techniques.
We are investigating techniques to extract high-level events from low-level RFID data. For example, how can we detect a database meeting event from a spatio-temporal combination of RFID tag sightings?
Monitoring applications enable users to continuously observe the current state of a system (e.g., car traffic conditions, a computer network, a cluster of servers). Although the current state of the system is the focus of monitoring applications, when events of interest occur, historical information is usually necessary to explain these events and determine appropriate responses. In this project, we are exploring techniques for complementing real-time monitoring information with different types of historical data.
Today, network intrusion detection systems (NIDSs) use custom solutions to log historical network flows and support forensic analysis by network administrators. In this project, we challenge the assumption that custom databases are needed and investigate how relational databases can support interactive network forensic analysis.
Here's the complete list.
If you are looking for our regular database group meetings and talks, they are under CSE 591D
Hosted talks at the Northwest Database Symposium (NWDS)