the cougar approach to in-network query processing in sensor networks
DESCRIPTION
These slides presents the research paper: "The cougar approach to in-network query processing in sensor networks" by Yong Yao and Johannes Gehrke. This presentation was done in a research seminar at the University of Colombo, School of Computing by the uploader.TRANSCRIPT
Presented By: Dilini A. MuthumalaSupervised By: Dr. Jeevani Goonetillake
The Cougar Approach to In-Network Query Processing
in Sensor Networks
Authors
Yong Yao
• Software Engineer at Google
• Ph.D., Computer ScienceCornell University (2000 – 2007)
• Research Interests– Databases– Sensor Networks– Distributed Systems
Johannes Gehrke
• University ProfessorDepartment of Computer Science
Cornell University
• Research Interests
–Scalability in Computer Games and Simulations–Data Privacy
–Data Mining
Motivation
• “Database Abstraction Layer” for Sensor Networks
• Most popular sensor data management middleware
• Introduces Database Abstraction Layer Concept
• Cited by 1185 (source: Google Scholar)No. of citations
Year
Presentation Outline
• Introduction
• Database Abstraction Layer
• Architecture
• Research Problems
• Conclusion
Introduction
Wireless Sensor Network (WSN)
Limitations
• Communication
• Power Consumption
• Computation
• Uncertainty in Sensor Readings
WSN Applications
• Smart Buildings, Smart Homes
WSN Applications
• Wild Life Monitoring
WSN Applications
• Monitoring Vineyards
Future of WSN
Is Johannes in his
office?
Internet
Future of WSN
Temperature
Humidity
Light
Motivation
1) Declarative queries are suited for WSN interaction
SELECT TempFROM sensors
Complex Network
Motivation
2) Increasing network lifetime is the major goal of any WSN application
WSNData Repository for
offline analysis
Database Abstraction Layer
Database Abstraction Layer
WSNWSNBase Station
SELECT Temp, Humid, NodeIDFROM sensorsSAMPLE PERIOD 5s
Node ID Temperature Humidity
1 127 44
2 119 47
3 120 45
4 123 40
5 120 46
Database Abstraction Layer
• Local computations are much cheaper than communication– Pushing partial computations out into the
network
Database Abstraction Layer
• Retrieves data only upon user demand
• No offline data storage
• Energy Efficient
Architecture
Architecture - Overview
Query OptimizerQuery Proxy Layer
Query Proxy Layer
Application Layer
Query Proxy Layer
Routing Layer
Other Layers
Query Optimizer• Generates “Query Processing Plans”
• Refers to– Catalog Information– Query Specification
• Specifies– Data Flow between sensors– Computation Plan
• Finally, plan is disseminated to all sensors
Example
User Query
“Notify when the average temperature
exceeds 35 °C”
WSNWSN
Query Optimizer
“Notify when the average temperature
exceeds 35 °C”
Query Optimizer
Query Plan
Query Plan (QP)
• Designates the Leader node– Where average value will be finalized
Query Plan
Leader Node
Query Plan (QP)
• Two computation plansi. Leader Node
ii. Non-Leader Nodes
Query Plan
QP for Non-Leader Node
Non-Leader Node
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
1
Temperature = 38 °C
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
2
Temperature = 38 °C
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
2
Temperature = 38 °C
AVG Temperature = 35 °CContributor Count = 1AVG Temperature = 36 °CContributor Count = 1
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
Temperature = 38 °C
AVG Temperature = 35 °CContributor Count = 1AVG Temperature = 36 °CContributor Count = 1
In-Network Aggregation
Total Temperature = 35*1 + 36*1 + 38
= 109
No of Contributors = 3
AVG Temperature = 109 / 3
= 36.33
Temperature = 38 °C
AVG Temperature = 35 °CContributor Count = 1
AVG Temperature = 36 °CContributor Count = 1
AVG Temperature = 36.33 °CContributor Count = 3
QP for Non-Leader Node
Sensor Scan
Network Interface
In-network Aggregation
AVG Temperature = 36.33 °CContributor Count = 3
Towards the Leader
QP for Leader Node
Leader Node
QP for Leader Node
Aggregate Operator (AVG)
Network Interface
Select AVG > threshold
Towards the Leader
Average Value
Partially aggregated results
QP for Leader Node
Aggregate Operator (AVG)
Network Interface
Select AVG > threshold
Towards the Leader
Average Value
Partially aggregated results
1
QP for Leader Node
Leader Node
AVG Temperature = 36.33 °CContributor Count = 3
AVG Temperature = 39 °CContributor Count = 2
QP for Leader Node
Aggregate Operator (AVG)
Network Interface
Select AVG > threshold
Towards the Leader
Average Value
Partially aggregated results
AVG Temperature = 36.33 °CContributor Count = 3
AVG Temperature = 39 °CContributor Count = 2
Aggregate Operator
Total Temperature = 39*2 + 36.33*3 = 186.99
No of Contributors = 5
AVG Temperature = 186.99 / 5
= 37.40
AVG Temperature = 36.33 °CContributor Count = 3
AVG Temperature = 39 °CContributor Count = 2
AVG Temperature = 37.40 °C
QP for Leader Node
Aggregate Operator (AVG)
Network Interface
Select AVG > threshold
Towards the Leader
Average Value
Partially aggregated results
AVG Temperature = 37.40 °C
QP for Leader Node
Aggregate Operator (AVG)
Network Interface
Select AVG > threshold
Towards the Leader
Average Value
Partially aggregated results
AVG Temperature = 37.40 °C
Threshold = 35 °C“Notify when
the average temperature exceeds 35 °C”
WSNWSN
User Query Result
ALERT!Temperature exceeds 35 °C
Research Problems
1. Aggregation
• Most popular computation and communication pattern
• Two important issues– Leader Selection– Data Delivery
Leader Selection
Requirements for the policyi. Dynamically-maintained Leader
ii. Physically advantageous location
Leader Selection
Requirements for the policyi. Dynamically-maintained Leader
ii. Physically advantageous location
Data Delivery
“How should the data be delivered from source nodes to the leader?”
– Send all data to leader?– Should intermediate nodes participate?
2. Query Language
“What types of queries should be supported?”
3. Query Optimization
• Cost of query plan has changed
• Energy should be the focus
• Reactive to changes in catalog information– Changes in topology– Power level at sensor nodes
4. Catalog Management
• Maintained at the server
• Provides Meta Data about the network
• Question: What is the best way to main the catalog?
5. Multi-Query Optimization
• Occurs when the WSN is shared
• Users may pose similar queries
• Share common data among the users
Conclusion
• Interacting with a WSN is made easy
• Database Abstraction layer provides– Friendly Interface
– Efficient scheme to reduce energy consumption
• Research problems need to be carefully addressed
My Views on the Paper
• Presents a concept
• Easy-to-understand
• Flow of the paper sometimes confuse the reader
Q & A