PERFORMANCE EVALUATION OF COMMON POWER ROUTING FOR AD-HOC NETWORK

1
PERFORMANCE EVALUATION OF COMMON POWER ROUTING
FOR AD-HOC NETWORK

• Zhan Liang
• Supervisor Prof. Sven-Gustav Häggman
• Instructor Researcher Boris Makarevitch
• Helsinki University of Technology
• Communications Laboratory
• 18th, May, 2004

2
Contents

• Background
• Objectives
• Introduction
• Implementation
• Evaluation of COMPOW
• Conclusion
• Future Work

3
What is Ad-hoc

• A local area network, or some small networks,
  parts are time-limited, and only usable for the
  duration of a communication session
• The routers are free to move randomly, organize
  themselves arbitrarily
• The wireless topology vary rapidly and
  unpredictably

4
Background

• Many power control methods are designed and
  implemented over Ad-hoc networks routing
  protocols (CLUSTERPOW, COMPOW, MINPOW, etc.)
• Few evaluation reports on the power control
  methods can be found

5
Why power control methods?

• A big effect on improving network capacity
• A higher transmit power
• a higher range and a higher signal-to-noise ratio
  to the receiver
• more interference to the adjacent nodes.
• Power control ? reduce the interfering nodes ?
  improve the capacity
• Energy Savings

6
Objectives

• To implement a common power control method
  (COMPOW) over one Ad-hoc networks routing
  protocol, AODV
• To evaluate this power control method

7
Introduction

• Ad-hoc routing protocols
• Power control methods

8
Ad-hoc routing protocols(1)

• Table-driven all the nodes know the routing
  information of the whole network
• Source-initiated routes are established only
  when the source nodes require them

9
Ad-hoc routing protocols(2)
10
Table-driven routing protocols Destination-Sequen
ced Distance-Vector (DSDV)

• To find the shortest paths, the least hops
• A routing table where all the routing information
  is stored

11
Source-initiated routing protocols(1)Dynamic
Source Routing (DSR)

• A route cache to cache the known routes to the
  destinations
• Main routing functions
• Route discovery
• Route maintenance

12
Source-initiated routing protocols(2)Ad-hoc
On-Demand Distance Vector (AODV) (1)

• A combination of both DSR and DSDV protocols
• The basic route-discovery and route-maintenance
  of DSR,
• The hop-by-hop routing, sequence numbers and
  beacons of DSDV

13
Source-initiated routing protocols(3)Ad-hoc
On-Demand Distance Vector (AODV) (2)

• Route discovery

14
Power control methods(1)

• COMPOW (COMmon POWer) control method
• CLUSTERPOW (CLUSTERing POWer) control method
• MINPOW (MINimum POWer) control method

15
Power control methods(2) COMPOW

• All the nodes use the same power level, the
  lowest power level at which the network is
  connected

16
Power control methods(3) CLUSTERPOW

• To separate nodes into several different clusters

17
Power control methods(3) MINPOW

• Each node chooses the transmit power level

18
Implementation of COMPOW(1)Simulation
Assumptions (1)

• Simulation Environment NS2
• Network card CISCO Aironet 350
• The channel is bi-directional link
• The free space loss with two ray ground
  reflection model

19
Implementation of COMPOW(2)Simulation
Assumptions (2)

• The antennas are omni directional (same gain and
  attenuation in all horizontal directions)
• The MAC layer protocol IEEE 802.11b

20
Implementation of COMPOW(3)COMPOW over AODV
Route Discovery procedure
21
Implementation of COMPOW(4)Architecture
22
Implementation of COMPOW(5)Functions included in
Simulation

• Route Discovery
• Route Maintenance
• Route Release
• Route Error handle

23
Evaluation of COMPOWTesting Scenarios

• Scenario 1 10 fixed nodes, 10 pairs of
  connection, 100 seconds, 250 m2
• Scenario 2 25 fixed nodes, 25 pairs of
  connection, 100 seconds, 625 m2
• Scenario 3 25 mobile nodes, 25 pairs of
  connection, 1000 seconds, 10001000 m2

24
ResultsThroughput vs. Load for fixed nodes (TCP)
25
ResultsThroughput vs. Load for fixed nodes (UDP)
26
ResultsEnergy Consumption vs. Load for fixed
nodes (TCP)
27
ResultsEnergy Consumption vs. Load for fixed
nodes (UDP)
28
ResultsThroughput vs. Load for mobile nodes
29
ResultsEnergy Consumption vs. Load for mobile
nodes
30
Conclusions

• A network transmitting packets by TCP COMPOW
  performs good
• A network transmitting packets by UDP the
  lifetime of the COMPOW network may be even
  shorter than that of the network without using
  power control methods

31
Future works

• More complicated scenarios test ? acquire a
  complete evaluation
• Non-uniform load generation environment
• Other Ad-hoc routing protocols ? a more complete
  evaluation of COMPOW

32
Q A

• Thank you for your attention!