Geographic Routing without Location Information

1. Geographic Routing without Location Information Yu-Min Tseng

2. Routing in Wireless Network • Distance vector • DSDV • On-demand • DSR, TORA, AODV • Discovers & caches routes on demand • Geographic • GPSR, DREAM, GRID • GPSR – scales very well

3. What’s the problem? • No address aggregation • Except geographic routing, all other approaches require O(N) state per node • Routing by coordinates is a good way to avoid O(N) per-node routing state

4. Why geographic routing without location information? • GPS takes power, doesn’t work indoors • Obstacles, no-ideal radios • Coordinates computed will reflect true connectivity & not the geographic locations of the nodes

5. Geographic routing • Choose coordinates for nodes • Greedy routing • Proceed closer to destination at each hop • How to deal with voids? • “Addresses” of nodes keep changing as they move • Need a lookup service for the current location of a node

6. Outline • Perimeter nodes & their locations are known • Perimeter nodes are known but their locations are not known • Nothing is known about the perimeter • Performance evaluation

7. Rubber Bands • Iterative process for picking coordinates for a node • Some nodes along with periphery of the network know their correct (relative) locations and are fixed • Other nodes compute coordinates by relaxation

8. Every node moves to the average of its neighbors coordinates at each step

9. Evaluation • Success Rate • Using purely greedy routing • Without dead-end avoidance techniques • Average Path Length • Average number of hops taken along the path

10. True geographic coordinate 3200 nodes uniformly spread, 64 perimeter nodes, Radio range is 8 unit. Each node have avg 16 neighbors.

11. Perimeter nodes are known (10 iterations) All non-perimeter nodes start with same initial coordinates

12. Perimeter nodes are known(100 iterations)

13. Perimeter nodes are known (1000 iterations) Success rate 0.993, average path length 17.1

14. True geographic coordinate Success rate 0.989, average path length 16.8

15. Resilience of rubber bands Success rate 0.99, average path length 17.1

16. Resilience of the rubber band Success rate 0.981, average path length 17.3

17. Rubber bands (cont’d) • We need a periodic heartbeat between neighbors so each node can maintain a list of its neighbors • We just send the current position of the node along with the heartbeat packet it broadcasts • Each time a heartbeat packet is received, we re-compute the coordinate

18. Outline • Perimeter nodes & their locations are known • Perimeter nodes are known but their locations are not known • Nothing is known about the perimeter • Performance evaluation

19. Perimeter coordinate algorithm • Perimeter coordinate algorithm • Perimeter nodes broadcast a HELLO message to network, discover the distance to all other perimeter node  perimeter vector • Perimeter nodes broadcast their perimeter vector to network • Every perimeter node use a triangulation algorithm to compute the coordinates of all perimeter nodes • Perimeter nodes flood O(N2) overhead

20. Perimeter coordinate algorithm (cont’d) • Measured_dist(I, j) • Distance between node i & j in perimeter vector • Dist(I, j) • Geometric distance between the virtual coordinates of node I & j • For minimizing the potential energy consumed, make the coordinate proportional to the hop count distance

21. Bootstrapping beacons • Two designated bootstrap beacons flood with HELLO messages • Perimeter nodes compute coordinates with bootstrap nodes • Compute the center of gravity (CG) • CG forms the origin • First bootstrap node defines x axis • Second bootstrap node define y axis

22. After 1 iteration, success rate 0.992, average path length 17.2 After 10 iterations, success rate 0.994, average path length 17.2

23. Outline • Perimeter nodes & their locations are known • Perimeter nodes are known but their locations are not known • Nothing is known about the perimeter • Performance evaluation

24. Perimeter node criterion • If a node is the farthest away, among all its 2-hop neighbors from bootstrap nodes, then the node decides that it is on the perimeter • A designated bootstrap node periodically broadcasts, nodes periodically re-assess whether they lie on the perimeter or not • Need a leader election protocol to deal with failure of this bootstrap node • Overhead doesn’t depend on N

26. Coordinate projection on circle • Origin is the CG of perimeter nodes • Radius is the average distance of perimeter nodes from CG • Virtual circle makes it easier to support mobility • If a node concludes that it is on the perimeter, it projects itself on circle & stays fixed

27. After 10 iterations, success rate 0.996, average path length 17.3

28. Outline • Perimeter nodes & their locations are known • Perimeter nodes are known but their locations are not known • Nothing is known about the perimeter • Performance evaluation

29. Experiment design • Event driven packet level simulator • Doesn’t model application traffic or collisions • Scales to 3200 nodes with packet events & 12800 nodes without events • 3200 nodes distributed randomly in a 200x200 square. Radio range is 8, density is held constant while scaling up

30. Success rate of greedy routing

31. Weird shapes

32. Different network densities

33. When greedy routing fails

34. Obstacles