YOU ARE DOWNLOADING DOCUMENT

Please tick the box to continue:

Transcript
Page 1: file anaylsis

A Distance RoutingEffect Algorithm for Mobility

(DREAM)*

Stefano BasagniIrnrich ChlamtacViolet R. SyrotiukBarry A. Woodward

Page 2: file anaylsis

Outline

• Introduction• DREAM -Dissemination of Location Information -A model of DREAM -DREAM procedure • Simulations and Results• Conclusion

Page 3: file anaylsis

Introduction

• Old problem for Ad hoc network routing: -Proactive: it corresponds to a next hop

table lookup, sequence is not explicit; -Reactive: the movement of any node in the

sequence renders the path invalid.• A new definition of routing table entry is

needed

Page 4: file anaylsis

Dissemination of Location Information

(DREAM)• Each node transmits control messages bearing its

current location to all the others. (e.g. geographic coordinates; obtained by the use of GPS [7])

• The frequency with which these control messages is determined by:

-distance effect -mobility rate

Page 5: file anaylsis
Page 6: file anaylsis

Dissemination of Location Information (control message)

• We assign each control packet a life time • A majority of the packets have a “short” life time: short lived packets are sent at high frequency, and “die” after they have traveled.• Long lived packets, sent less frequently, travel farther through the network.

Page 7: file anaylsis

A Model for DREAM

• S sends a message to node R, it refers to its LT (Location Table) in order to retrieve location information about R.

• S selects from among its neighbors those nodes that are in the direction of R

• It is guaranteed that R can be found with a given probability p, 0< p <1, following result in that direction.

Page 8: file anaylsis
Page 9: file anaylsis

A Model for DREAM

• The time interval from t0 to tl, tl > t0• x = (tl – t0)v• Node R, whose speed is v, cannot be anywhere outside the circle C• one hop neighbors those nodes A, direction A. lies within the range [θ- α, θ+ α] • Angle α must be chosen in a way that the probability of finding R in the sector S is at least p.• we want to find a minimum value for α

Page 10: file anaylsis

A Model for DREAM (method of finding α)-1

Page 11: file anaylsis

A Model for DREAM (method of finding α)-2

Page 12: file anaylsis

Distance Routing Effect Algorithm

for Mobility (DREAM procedure)

Page 13: file anaylsis

Distance Routing Effect Algorithm

for Mobility (DREAM procedure)

Page 14: file anaylsis

Distance Routing Effect Algorithmfor Mobility (DREAM procedure:Recovery)

• Its actual implementation may vary, depending on the characteristic of the network.• For instance, flooding

Page 15: file anaylsis

Simulations Results

• Simulated our DREAM protocol using MAISIE [1]

• Placing n = 30 nodes randomly on a grid of size 100 x 100.

• we assume that each node has the same speed V• given in grid units per 100 ticks of the simulation

clock• average end-to-end delay

Page 16: file anaylsis
Page 17: file anaylsis
Page 18: file anaylsis
Page 19: file anaylsis

Conclusions

• Simulation results showed that with over 80% probability this method can find a route to a given node. (if any exists)• The average end-to-end delays with respect to the DSR reactive protocol are lower.• DREAM protocol provides loopfree routes, and is robust in providing multiple routes.