Aggregating and Querying Geo-Streams : From static sensors to Evolving Phenomena

Aggregating and Querying Geo-Streams:

From static sensorsto Evolving Phenomena

Chengyang ZhangComputer Science Department

University of North Texas

06/26/08 CRI Group Meeting

• Refers to data emitted by geo-referenced sensors and sensor networks, such as water level, vehicle speed

• Queries are often interested in moving/evolving phenomena, such as flood, traffic jam

• Need to bridge the gap between the point observations and continuousphenomena in REAL time.

Geo-Stream

CRI Group Meeting

From Point Observations to Moving Regions

CRI Group Meeting

Challenges

• Supporting spatio-temporal stream data types

• Allow geo-streams to be grouped according to spatial proximity and then elevated into streaming regions

• Query Processing and Optimization

CRI Group MeetingMarch 26, 2008 5SMU Talk

1. Stream Data Types

CRI Group MeetingMarch 26, 2008 6SMU Talk

2. Generalize Aggregate Functions

CRI Group MeetingMarch 26, 2008 7SMU Talk

Generalize Aggregate Functions

CRI Group MeetingMarch 26, 2008 8SMU Talk

3. Query Optimization

CRI Group MeetingMarch 26, 2008 9SMU Talk

Semantics-based Optimization

CRI Group MeetingMarch 26, 2008 10SMU Talk

Incremental Optimization

CRI Group MeetingMarch 26, 2008 11SMU Talk

Experimental Results

CRI Group MeetingMarch 26, 2008 12SMU Talk

Experimental Results

06/26/08 CRI Group Meeting

Questions?