1. Efficient Geographic Routing in Multihop Wireless Networks Seungjoon Lee*, Bobby Bhattacharjee*, and Suman Banerjee** *Department of Computer Science University of Maryland **Department of Computer Sciences University of Wisconsin-Madison Proceedings of the 6th ACM international symposium on Mobile ad hoc networking and computing (MobiHoc '05) Chien-Ku Lai

2. Outline Introduction New Link Metric for Geographic Routing Link Cost Types and Estimation Simulation Conclusions and Future Work

3. Introduction - Geographic Routing (Position-based Routing) This kind of routings uses location information for packet delivery Neighbors locally exchange location information Neither route establishment nor per-destination state is required

4. Introduction - Normalized Advance (NADV) Instead of the neighbor closest to the destination, NADV lets users select the neighbor with the best trade-off between link cost and proximity

5. Introduction - about this paper For the effective use of NADV, this work presents techniques for efficient and adaptive link cost estimation Providing multiple techniques thus enabling nodes to choose the best scheme for the current network and system setting

6. New Link Metric for Geographic Routing 1. Background 2. Normalized Advance

7. Background Link cost  the power consumption required for a packet transmission over the link Link metric  “degree of preference” in path selection

8. Background (cont.) In many geographic routing protocols  The current node S greedily selects the neighbor that is closest to destination T

9. Background - Goal To gain as large advance as possible for fast and efficient packet delivery To balance the trade-off, so that we can select a neighbor with both large advance and good link quality

10. Normalized Advance (NADV)

11. Link Cost Types and Estimation 1. Packet Error Rate (PER) 2. Delay 3. Power Consumption

12. Packet Error Rate (PER)

13. Using Probe Messages for PER Estimation Using Signal-to-Noise Ratio for PER Estimation Neighborhood Monitoring for PER Estimation Self Monitoring for PER Estimation

14. Packet Error Rate (PER) - Using Probe Messages for PER Estimation

15. Packet Error Rate (PER) - Using Signal-to-Noise Ratio for PER Estimation Assuming an AWGN (Additive White Gaussian Noise) channel, in the case of BPSK (Binary Phase Shift Keying), the bit error rate is given by ： the received power ： the channel bandwidth ： the noise power ： the transmission bit rate ： the complementary error function

16. Packet Error Rate (PER) - Neighborhood Monitoring for PER Estimation In IEEE 802.11 networks  using the MAC sequence number A can count how many frames from neighbor B it has missed The quality of two directional links may differ

17. Packet Error Rate (PER) - Self Monitoring for PER Estimation Aging  multiply PERs of unused links by 0.9 every 30 seconds

18. Delay Two types of link delay  medium time  total delay – future work

19. Power Consumption

20. Simulation 1. Model 2. Results

21. Simulation Model Simulator ： ns-2 Deployment ： uniform Region ： 1000m x 1000m Nodes ： 100 Maximum transmission range ： 250m

22. Simulation Results - Number of transmissions/delivery

23. Simulation Results - Average Path Length

24. Simulation Results - Latency

25. Simulation Results - Using Delay as Link Cost

26. Simulation Results - Power Consumption

27. Conclusions and Future Work

28. Conclusions This work has introduced NADV as link metric for geographic routing Geographic routing with NADV provides an adaptive routing strategy  is general  can be used for various link cost types This work presented techniques for link cost estimation NADV also finds paths whose cost is close to the optimum

29. Future Work To design a link cost model that balances multiple cost criteria To implement the NADV framework on real testbeds and evaluate the performance in practice

30. Question? Thank you.