Efficiently searchable overlay for multimedia sensor storage
Surendar Chandra
Notre Dame University
Abstract
Traditional
wireless sensor network deployments use a wide variety of low fidelity sensors
that capture environmental parameters such as temperature, humidity, light,
motion etc. to detect and respond to potential threats. Our work focuses on
widely deploying high fidelity multimedia sensors and storing the captured
streams in in-situ storage. Smarter sensors analyze the stored streams to detect
emerging threats; easily missed by simpler real time algorithms. Storage bricks
are freely deployed alongside multimedia sensors to spatially localize the
streams and allow for incrementally scalable storage. Our research focus is on
self-managing and self-organizing mechanisms for a scalable storage of the
sensed multimedia.
In this presentation, I will briefly outline some of the research
challenges for our multimedia storage. I will also outline our overlay
mechanisms that spatially localize the streams while maintaining good
connectivity. Our unstructured overlay allows for efficient searches that can
be leveraged to optimize for segment storage or retrieval. For a 10000 node
system, we show that our system can locate objects within four hops for replication
ratios as low as 0.05%.
Biography
Surendar Chandra is an assistant professor in the
department of Computer Science & Engineering at the University of Notre Dame. He spent 2000-2002 as a faculty member at the