Presentation is loading. Please wait.

Presentation is loading. Please wait.

Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Stochastic Characterization of Mobile Ad-hoc Networks John P. Mullen and Timothy I. Matis.

Similar presentations


Presentation on theme: "Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Stochastic Characterization of Mobile Ad-hoc Networks John P. Mullen and Timothy I. Matis."— Presentation transcript:

1 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Stochastic Characterization of Mobile Ad-hoc Networks John P. Mullen and Timothy I. Matis Center for Stochastic Modeling Department of Industrial Engineering New Mexico State University INFORMS 2004

2 2 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 What Are MANETS ?  A MANET is a mobile ad-hoc wireless communication network that is capable of autonomous operation  Each node is capable of transmitting, receiving, and routing packets of information.  The network has no fixed backbone  The nodes are able to enter, leave, and move around the network independently and randomly

3 3 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Mobile Ad Hoc Path Search Y X A B I G E F C D H

4 4 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Same MANET After a While Y X A B I G E F C D H H X I G F E D B A C Y

5 5 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Nutshell  MANET field performance differs greatly from simulation’s  Field & testbed performance is much poorer  Developing MANET protocols in the field is very difficult  Improving simulation fidelity increases the value of simulation in design.  Higher fidelity earlier in the design process leads to better designs  Research focus:  Significantly improve the fidelity of MANET simulations  Without significantly increasing  Simulation run time or  Modeling effort.  Research results  Up to an order of magnitude improvement in fidelity  Runtime increases are often insignificant, but generally less than 100%  Very little added modeling effort

6 6 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Overview  Multipath Fading and its impact on mobile ad hoc nets  The Stochastic Model  Objectives  Implementation  Validation  Demonstrations of the Model  Small Models  Impact of Short Retry Limit (SRL)  Comparing AODV and DSR  Large Models  AODV vs. DSR  AODV vs. DSR using GPS data  Impact of SRL on DSR  Summary, Conclusions and Further Work

7 7 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Shadowing and Fading  Shadowing  Is caused by objects absorbing part of the signal  Can be estimated by looking at the Line of Sight (LOS) path  Causes a random reduction in signal strength.  Fading  Is the result of the algebraic sum of signals from many paths  Because movement of any object in the vicinity can change the sum  Multipath fading is extremely difficult to model and predict  Would be very time consuming to simulate exactly  And would have little predictive value.  This phenomena causes:  Very rapid large-scale fluctuations in signal strength  Can cause the signal to be significantly lesser or greater than expected.

8 8 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Main causes of signal variation R T Shadowing Multipath

9 9 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Measured Received Signal Strength (from Neskovic 2000)

10 10 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Stochastic Variation Model  The Model  Given m p (d), the expectation of power at distance d  Rayleigh fading model of the instantaneous power, P(d)  Pr {P(d) ≤ p} = 1 – exp{-[p/m p (d)]}  Inverse transform of the Rayleigh fading model  P(d) = -m p (d)ln(1-r)

11 11 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Simulated vs. Real Power Actual Measurements Simulated Values

12 12 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Validation  Simulated reported field tests and compared results  K.-W. Chin, J. Judge, A. Williams, and R. Kermode, "Implementation experience with MANET routing protocols," ACM SIGCOMM Computer Communications Review, vol. 32, pp. 49 - 59, 2002.  I. D. Chakeres and E. M. Belding-Royer, "The Utility of Hello messages for determining link connectivity," Wireless Personal Multimedia Communications, vol. 2, pp. 504 - 508, 2003.  D. S. J. D. Couto, D. Aguayo, J. Bicket, and R. Morris, "A High Throughput Path Metric for MultiHop Wireless Routing," presented at MobiCom '03, San Diego, California, USA, 2003.  S. Desilva and S. Das, "Experimental evaluation of a wireless ad hoc network," 2000.  Simulations with  Standard non-fading model were exceedingly optimistic  Proposed fading model were very much more realistic.

13 13 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Impact of Multipath Fading on MANETs  How does it affect MANETs?  Unnecessary route searches  Selection of false routes

14 14 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Impact of Multipath Fading On MANETs Nominal Range (r 0 ) OK Stub Cellular Dropped Packets Fading margin False Routes OK

15 15 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Impact of Multiple Retries on MANETs

16 16 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 The MANET fading Trade-off Protocol Improve Reliability On Good Routes Increase Risk of Selecting Bad Routes MANET: Nominal range is a matter of balance. Most Wireless: Nominal range is a matter of design.

17 17 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Demonstrations  Small Models (validations of field tests)  Scenario 1 – Performance vs. distance.  Used for the two cases above  Scenario 2 – Routing Test  Focus mainly on fading effects  Models:  Fading vs. nonfading simulations of AODV  DSR vs. AODV with fading model  Large models (exploration)  Scenario 3 – 24 nodes.  Also consider other effects, such as interference  Models: Fading and non-fading versions of  AODV vs. DSR  AODV vs. DSR using GPS data  Impact of SRL on DSR

18 18 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Scenario 2: Routing test (from Chin et. al., 2002) 10 pps 0.5 m/s r 0 = 39m

19 19 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Sc 2: Fading vs. Nonfading: AODV Notes: Default values for AODV SRL = 7

20 20 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Sc 2: AODV vs. DSR Notes: Default protocol values SRL = 7 Nonfading model shows no difference

21 21 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Scenario 3: Larger Scale Test Features: More nodes (24) Random r-t pairs Interference Higher loads

22 22 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Mean Throughput: AODV vs. DSR Notes: Default protocol values SRL = 7

23 23 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Mean Delay: AODV vs. DSR Notes: Default protocol values SRL = 7

24 24 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Using GPS data 3 2 1 B A r0r0 Use GPS to block unreliable routes

25 25 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Impact of GPS Without GPS With GPS

26 26 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Mean Throughput: Impact of SRL on DSR Notes: Default protocol values

27 27 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Mean Delay: Impact of SRL on DSR Notes: Default protocol values

28 28 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Execution Time in Scenario 3 (Virtually no differences in Scenarios 1 & 2)

29 29 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Summary  Non fading model  Overestimates field performance  Is very insensitive to all the contrasts shown here and more.  Fading model  Provides more realistic estimates  Better predicts impacts of protocol and parameter changes  Shows promise of new techniques.  Requires little or no additional modeling  Has little impact on execution time  (Alternative is a testbed or a field trial)

30 30 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Conclusions  Multipath Fading  has a great impact on mobile ad hoc nets  Including its effects in simulation  greatly improves fidelity  Stochastic Modeling of Multipath Fading  Is a practical way to include the impact of fading  Minor modifications to code (in OPNET, at least)  Without great increases in  Modeling effort or  Execution time

31 31 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Future Work  More Fading Models Rayleigh  Ricean  Nakagami  Other significant RF effects  e.g. exponential decay factor  Better user interface  Allow selection of models & parameters without need to recompile.  Validation Replicating published studies  Set up own testbed and field trials  Better modeling of fading impacts  Hello vs. control vs. data packet results  Other significant measurable elements.

32 32 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Acknowledgements  OPNET Technologies  Software license research grant  Technical assistance  Center for Stochastic Modeling  Technical resources  Klipsch School of Electrical and Computer Engineering  Dr. Steve Horan  Dr. Hong Huang (also CSM member)

33 33 INFORMS 2004 Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Final Questions?            


Download ppt "Primary Author: J.P. Mullen, Presented by: T.I. Matis, 10/26/2004 Stochastic Characterization of Mobile Ad-hoc Networks John P. Mullen and Timothy I. Matis."

Similar presentations


Ads by Google