1. TTDD: A Two-tier Data Dissemination Model for Large-scale Wireless Sensor Networks
Haiyun Luo, Fan Ye, Jerry Cheng, Songwu Lu, Lixia Zhang
UCLA Computer Science Department
Presenter: Tylor

2. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

3. Introduction Sensor network applications Basic problem Source Sink
Environmental/eco-system monitoring
Health care
Elevator, Presence detection, … etc
Basic problem
Source
Sink
But how? ?

4. Introduction Design Consideration New challenge
Energy efficiency
Reliability
Short delay
New challenge
Mobility
Two-tier data dissemination (TTDD)
Efficient data dissemination in large-scale wireless sensor networks with sink mobility

5. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

6. Two-tier Data Dissemination
Assumption
Static sensor nodes (except the sinks)
Location-aware
Each node is aware of its own location
Mission-aware

7. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

8. Grid Construction
Once detection of a stimulus, source divides the network into a grid of cells

9. Grid Construction Determination of grid points
α : cell size
(X, Y) : location of source
(Xi, Yi) : location of grid points
Source sends a data announcement message to four adjacent grid points using simple greedy geographic forwarding (GF)

10. Grid Construction
A node is closer to the grid point than all its neighbors
If the distance is less than α/ 2
this node becomes dissemination node
Else
drop the message
End
Dissemination node stores some information
The location of the grid point it is serving
Upstream dissemination node’s location
The grid is built on a per-source basis

11. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

12. Two-tier Query and Data Forwarding
The lower tier is within the local grid of the sink’s current location
The higher tier is made of the dissemination nodes
Sink floods the query locally to find a (immediate) dissemination node

13. Query Forwarding and Data Forwarding
Each dissemination node stores the location of the downstream dissemination node

14. Trajectory Forwarding
The primary agent (PA)
The immediate agent (IA)
Initially PA and IA are the same sensor nodes

15. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

16. Grid maintenance The failure of sensor nodes
Upstream information duplication
Dissemination nodes
Timeout at the sink
Immediate dissemination nodes
Primary agent
Immediate agent
There is also a timer for the grid state

17. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

18. Simulation Setups ns-2.1b8a 4 sources, 4 sinks Mobility model α (m)
Random way point
Sink’s speed 10 m/s
5-sec pause time
Local flooding range
1.3 α
Simulation time (s)
200
Network size (m)
2000
MAC
DCF
Number of nodes
Data rate (event/s) 1
Transmit energy (W)
0.66
Receive energy (W)
0.395
Idle energy (W)
0.035
α (m)
600

19. Reachability v.s. Sink Mobility (4 srcs, 4 sinks)
The capability of handling mobility

20. Increasing energy consumption
Sink’s local flooding queries
The high-tier grid forwarding effectively localizes the impact of sink mobility

21. Reachability v.s. Node Failures (4 srcs, sinks, 10m/s)
15% randomly-chosen nodes fail at 20 seconds

22. Reduced data packet delivery
Repair of failed dissemination nodes

23. TTDD v.s DD
Static sinks

24. Reachability v.s. Sink # No dissemination node serves for the sink
Collisions

25. Energy v.s. Sink # Less energy consumption when more than 2 srcs
Location estimation
High data rate
Less energy consumption in 1 or 2 srcs
Exploratory data
Query aggreagation

26. Delay v.s. Sink # Longer data path In cases of multiple sources
Data path overlap

27. Road Map Introduction TTDD Performance evaluation Discussion
Grid construction
Two-tier query and data forwarding
Grid maintenance
Performance evaluation
Discussion
Conclusion

28. Energy v.s. cell sizes (1 src, 1 sink, 10m/s)
The energy first decreases
Less energy is required to build a grid with larger cell size
The energy increases later
Local query flooding consumes more energy in larger cells
The decision of α depends on the mobility patterns of the sink

29. Conclusion
TTDD
Efficient data dissemination is large-scale sensor fields
Local floodings minimize the overall network load and the amount of energy to maintain data-forwarding paths
Sources and sinks cooperate to accomplish efficient data delivery to mobile sinks

30. Pros and cons
Pros
The idea is similar to GSM architecture which efficiently locates the mobility of user
Simple and work well
Cons
The grid structure which is built on a per-source basis may consume too much energy when the location of sources change frequently
The assumption of static nodes in the network and the poor failure recovery mechanism