1. Using Directional Antennas for Medium Access Control in Ad Hoc Networks MOBICOM 2002 R. Roy Choudhury et al. 2002.10.16 Presented by Hyeeun Choi

2. 2/26 Contents  Introduction  Related Works  Preliminaries  Basic Directional MAC (DMAC) Protocol  Multi-Hop RTS MAC (MMAC)  Performance Evaluation  Future Work  Conclusion

3. 3/26 Introduction  The Problem of utilizing directional Antennas to improve the performance of ad hoc networks is non-trivial  Pros  Higher gain (Reduced interference)  Spatial Reuse  Cons  Potential possibility to interfere with communications taking place far away

4. 4/26 Omni-directional Antennas S D A B Silenced Node C

5. 5/26 Directional Antennas S D A B C Not possible using Omni

6. 6/26 Related Works  MAC Proposals differ based on  How RTS/CTS transmitted (omni, directional)  Transmission range of directional antennas  Channel access schemes  Omni or directional NAVs  Gain of directional antennas is equal to the gain of omni-directional antennas

7. 7/26 Preliminaries (1/2)  Antenna Model  Two Operation modes : Omni & Directional  Omni Mode:  Omni Gain = Go  Idle node stays in Omni mode.  Directional Mode:  Capable of beamforming in specified direction  Directional Gain = Gd (Gd > Go)

8. 8/26 Preliminaries (2/2)  IEEE 802.11 IEEE 802.11 DCF – RTS/CTS access scheme Physical Carrier Sense Physical Carrier Sensing Virtual Carrier Sensing

9. 9/26 Problem Formulation  Using directional antennas  Spatial reuse  Possible to carry out multiple simultaneous transmissions in the same neighborhood  Higher gain  Greater transmission range than omni-directional  Two distant nodes can communicate with a single hop  Routes with fewer hops

10. 10/26 Basic DMAC Protocol (1/2)  Channel Reservation  A node listens omni-directionally when idle  Sender transmits Directional-RTS (DRTS) using specified transceiver profile  Physical carrier sense  Virtual carrier sense with Directional NAV  RTS received in Omni mode (only DO links used)  Receiver sends Directional-CTS (DCTS)  DATA,ACK transmitted and received directionally

11. 11/26 Basic DMAC Protocol (2/2)  Directional NAV (DNAV) Table  Tables that keeps track of the directions towards which node must not initiate a transmission E H B 2*ß ε θ ε = 2ß + Θ If Θ> 0, New transmission can be initiated DNAV C CTS RTS

12. 12/26 Problems with Basic DMAC (1/4)  Hidden Terminal Problems due to asymmetry in gain  A does not get RTS/CTS from C/B C A B Data RTS

13. 13/26 Problems with Basic DMAC (2/4)  Hidden Terminal Problems due to unheard RTS/CTS C B D A

14. 14/26 Problems with Basic DMAC (3/4)  Shape of Silence Regions Region of interference for directional transmission Region of interference for omnidirectional transmission

15. 15/26 Problems with Basic DMAC (4/4)  Deafness RTS AB X Z DATA X does not know node A is busy. X keeps transmitting RTSs to node A

16. 16/26 MMAC Protocol (1/3)  Attempts to exploit the extended transmission range  Make Use of DD Links  Direction-Direction (DD) Neighbor C A A and C can communication each other directly

17. 17/26 MMAC Protocol (2/3)  Protocol Description : Multi-Hop RTS  Based on Basic DMAC protocol D R G S T B A C F DO neighbors DD neighbors RTS DATA

18. 18/26 MMAC Protocol (3/3)  Channel Reservation  Send Forwarding RTS with Profile of node F R G S T B C Fowarding RTS DATA A FD

19. 19/26 Performance Evaluation (1/6)  Simulation Environment  Qualnet simulator 2.6.1  Beamwidth :45 degrees  Main-lobe Gain : 10 dBi  802.11 transmission range : 250meters  DD transmission range : 900m approx  Two way propagation model  Mobility : none

20. 20/26 Performance Evaluation (2/6) A B C D E F AB CD E F High Spatial Reuse Aggregate Throughput (Kbps) IEEE 802.11 : 1189.73 Basic DMAC : 2704.18 High Directional Interference Hidden terminal Problem Aggregate Throughput (Kbps) IEEE 802.11 : 1194.81 Basic DMAC : 1419.51

21. 21/26 Performance Evaluation (3/6) 150m  Aligned Routes

22. 22/26 Performance Evaluation (4/6)  Less aligned Routes

23. 23/26 Performance Evaluation (5/6)  Randomly Chosen Routes

24. 24/26 Performance Evaluation (6/6)  Random Topology

25. 25/26 Future Work  Design of directional MAC protocols that incorporate transmit power control  New protocols that rely less on the upper layers for beamforming information  Impact of directional antennas on the performance of routing protocols

26. 26/26 Conclusion  Directional MAC protocols show improvement in aggregate throughput and delay  But not always  Performance dependent on topology  Random topology aids directional communication  MMAC outperforms DMAC & 802.11  802.11 better in some scenarios

27. 27/26 Related Works  Nasipuri et al  Assume that the gain is equal to the gain of omni- directional antennas  Ko et al  A node may transmit in directions that do not interfere with ongoing transmission  Bandyopahyay et al  Present another MAC which uses additional messages to inform neighborhood nodes about ongoing communications  Takai et al  Directional virtual Carrier Sensing (DVCS)

28. 28/26 Extra Slides  Gain of Antenna  Used to quantify the directionality of an antenna  Relative power in one direction compared to an omni-directional antenna