Transcript Frank Vernon

Open problems
•
•
•
•
•
•
How can I acquire desired RT data?
How can I discover/access desired RT data?
How can I share/integrate the data?
Can I integrate dynamically?
Can I share with other disciplines?
Can I integrate seamlessly with data from other
disciplines
• Can I integrate with archived data?
Sensor Networks:
Next Generation Problems
SAMSI
Sensor Network Workshop
14-15 October 2003
Frank Vernon
Scripps Institution of Oceanography
University of California at San Diego
Sensor Networks Desires
• Timing
– Critical issue for any modal analysis or array
processing analysis
– Need to synchronize across all sensors
• Communications systems
– Power < 1W for remote sites
– IP
– Aggregate data rate ≥ 1 Mbit/sec
– Scalable for array apertures from
100s meters => megameters
What is needed for a real-time data
exchange systems:
• System should be “data neutral”
– Variable packet size
– All formatting at application level
• System should allow data query functions
• System should allow dynamic updates of instrument
metadata
– metadata not embedded in every data packet
• Packet descriptions should be available in real-time
– Self-describing
What is needed for a real-time data
exchange systems:
• System should be network transparent with respect to
read/write/local/remote access
• System should be error-free and robust
– Impervious to communications failures
– Impervious to computer shutdown-startup
– Generally means non-volatile buffers, autoreconnects, auto-restarts, etc.
• Interoperable with other real-time data exchange
systems and new real-time data sources
Design Goals for Future
Real-Time Sensor Networks
• capture, process, control for quality, and integrate
real-time data streaming from different sources-collected for different purposes, on different time and
spatial scales, and measured by different methods;
• make heterogeneities in platforms, physical location
and naming of resources, data formats and data
models, supported programming interfaces, and
query languages transparent to the user;
• adapt to new and changing user requirements for
data and data products;
Design Goals for Future
Real-Time Sensor Networks
• dynamically reconfigure with the addition or removal of
observational equipment and/or scientific
instrumentation
– self-healing topology
– decentralized resource registration
• provide Internet access to integrated data collections
along with visualization, data mining, analysis, and
modeling capabilities