A Hierarchical Scheme for Data Aggregation in Sensor networks

1. A Hierarchical Scheme for Data Aggregation in Sensor networks Bin Zhou*; Lek Heng Ngoh**; Bu Sung Lee*; Cheng Peng Fu*; *School of Computer Engineering Nanyang Technological University **Networking Department Institute for Infocomm Research IEEE International Conference Networks (ICON), 2004

2. Outline • Introduction • Related work • Hierarchical Scheme • Performance Evaluation • Conclusion

3. Introduction - background • Data transmission over sensor networks • Location-centric protocol • TTDD: Two-tier Data Dissemination in Large-scale Wireless Sensor Networks • Data –centric protocol • Directed Diffusion for Wireless Sensor Networking

4. Introduction - motivation and goals • Motivation • Hierarchical scheme with enhancement on Directed Diffusion • Goals • Save transmission energy • Higher reliability

5. Related work • Title : Directed Diffusion for Sensor Networks • Author : Chalermek Intanagonwiwat, Ramesh Govindan, Deborah Estrin • IEEE/ACM Transactions on Networking(2003 )

6. Related work – Directed Diffusion for Sensor Networks • Data-centric dissemination • Request driven • Sinks place requests as interests • Sources are eventually found and satisfy interests • Intermediate nodes route data toward sinks • Reinforcement-based adaptation to the best path • Repair

7. Interests and Gradients • Flooding Interest • Gradient specifies both a data rate and a direction in which to send events Source Gradient Interest Sink

8. Data Propagation • Data flows from the source to the sink along the gradient • Multipath delivery with probabilistic forwarding. • Multipath delivery with selective quality along different paths. Source Gradient Data Sink

9. Reinforcement • Reinforcement to single path delivery • Selects an empirically low delay path Source Gradient Data Reinforcement Sink

10. Multiple sources Source Gradient Data Source Reinforcement Sink

11. Repair Gradient Source Data Reinforcement Sink Source Sink

12. Related work – Directed Diffusion for Sensor Networks • Advantage • Provide robust, scalable and energy efficient sensor data dissemination mechanism • Shortcoming • Data aggregation is not optimal

13. Hierarchical scheme -overview • Hierarchy construction • Attribute establishment • Data aggregation • Data distribution

14. Hierarchical scheme - Hierarchy construction • Flood interest message over sensor networks • Each node (level N) knows about all its parents and children (level N+2). sink Level 0 1 1 Level 1 2 2 2 2 2 Level 2 3 3 source source Level 3 source source source source

15. Hierarchical scheme -Attribute establishment • When source has the appropriate data, it will send back the exploratory data. • Attribute : number of sources contained in the sub-tree rooted by the node 3 2 3 2 2 source source source

16. Hierarchical scheme - Attribute establishment • Reinforcement message : parent informs its children about its own attribute 3 2 3 2 2 source source source

17. Hierarchical scheme -Data aggregation • A source receives the reinforcement, it sends out the data. • Child chooses the “alive” parent whose attribute is best. 3 2 3 2 2 source source source

18. Sink 3 Sink 2 3 2 Sink 1 2 3 1 1 2 3 source Hierarchical scheme -Data distribution • With multiple sinks, each sink will have an individual hierarchical structure. • Every node stores the information of its parent and children for different sink interests.

19. ns-2 50-250 nodes are randomly spread in a 160X160 m square field. Radio range : 40m 5 sinks and 5 sources Interest rate of sink : 1 event per 5 seconds Exploratory data rate of source : 1 event per 50 seconds Data rate of source : 2 event/s Performance Evaluation

20. Performance Evaluation

21. Performance Evaluation

22. Performance Evaluation • 20% of the nodes fail randomly

23. Conclusion • Designed a hierarchical scheme for data aggregation in sensor networks • energy saving • transmission efficiency • higher reliability