Performance Evaluation of WLAN for
Mutual Interaction between Unicast
and Multicast Communication Session
Author: Aamir Mahmood
Supervisor: Prof. Riku Jäntti
Wireless Local Area Networks

IEEE802.11 family of standards

Maturity of standard, low cost infrastructure, operation
in unlicensed band

Simple standalone infrastructure

Extensions to existing networks (WiMax, TETRA)

Providing access for high speed data and multimedia
services

Growing interest in outdoor operation of IEEE802.11
WLAN for Real-Time Services

Primary objective - asynchronous (data) services

Case: A realistic network with multiple
technologies complementing each other

Requirement: Support for data and real-time services

Challenges:
wireless

Uncontrolled and unreliable propagation environment

Stringent Quality of Service (QoS) requirements

Performance and scalability constraints of MAC algorithm
Thesis Contribution

WLAN evaluation by simulations under mutual
interaction of unicast and multicast real-time sessions


Extending the isolating study of unicast and multicast sessions
Evaluation Steps:

Designing a prototype for simulator verification (PHY and MAC
parameters, backoff time distribution, collision probability)

Mobility with a proposed model

Optimal size of WLAN cell, suitable for unicast and multicast
sessions under TWO-RAY propagation model
Test bed and Simulator


Test environment concerns

Modifications in system parameters

Reliability and reproducibility of the results
Test bed


Open Source Components: operating system, WLAN adaptor drivers,
real-time traffic emulation software, traffic monitoring and sniffing
Simulator

Qualnet 4.0 – Reliable and comprehensive modeling / simulation

Signal reception model

IEEE 802.11 PHY layer

Propagation model

IEEE 802.11 MAC layer
Prototype Design for Simulator
Verification - I

Maximum aggregate throughput
for two nodes with UDP flooding

Constant propagation
environment

Measure the collision probability

Two nodes sharing the equal
throughput
Node 1
30 dB
Var. Attenuator
40 dB
Access Point
(AP)
Monitoring Station
30 dB
RF Cable
Ethernet Cable
Node 2
Testbed vs Simulation Throughput - II
6.71
6
simulator
testbed
1.71
simulator
testbed
Throughput (Mbps)
Throughput (Mbps)
1.5
5
4
3
1.3
1.1
0.9
2
0.7
1
100
300
500
700
900
UDP Payload
@11Mbps
1100
1300
1472
0.6
100
300
500
700
900
UDP Payload
@2Mbps
1100
1300
1472
Collision Probability - III
*
Wavg 
1  W (1  p)(1  (2 p) 1  p
(2 W  1)( p  p ) 
.




1  pk 
2(1  2 p)
2
2

m
m
m
m
m
Nodes
Simulated collision
probability
Simulated average
backoff
Analytical average
backoff
2
0.0318
16.0603
16.0423
3
0.0605
16.8147
16.5878
4
0.0867
17.1745
17.1213
6
0.1439
18.2225
18.2999
8
0.1818
19.1257
18.9628
*H.L. Vu and T Sakurai, “Collision probability in saturated IEEE 802.11 networks”, ATNAC Australian
Telecommunication Networks and Applications Conference 2006
Group Mobility

The proposed mutual interaction of unicast and multicast
sessions is well-suited for simultaneous one-to-one and
group communication

The performance for the proposed joint flows is evaluated
under a proposed group mobility model

The model maps the mobility of public safety cooperative
activities
Proposed Mobility Model

The deployment of users towards
a randomly selected hotspot area
belonging to the cell

Uniform initial distribution

Speed of the user is proportional
to the distance from the
destination in the hotspot

The destination location of a user
in hotspot is also uniformly
distributed
Statistical Analysis - I
Rc = 1 units, Ro = 0.1 units, Vmin = 0.005 units/sec
Vmax = 0.015 units/sec
180
0.015
160
120
0.01
pdf
Velocity (units/s)
140
100
80
60
0.005
40
20
0
0
0.5
1
1.5
2
Distance (units)
Velocity as a function of distance
between the initial and final position
0
0.005 0.006 0.007 0.008 0.009
0.01
0.011 0.012 0.013 0.014 0.015
Velocity (units/s)
PDF of initial speed distribution
0.01
Ro=0.1
Ro=0.5
0.009
0.008
0.8
0.007
0.6
0.006
0.005
pdf
Instantaneous network speed (units/s)
Statistical Analysis - II
0.4
0.004
0.2
0.003
0
1
0.002
0.5
0.001
0
1
0.5
0
0
-0.5
20
40
60
80
100
120
Simulation time (s)
Instantaneous network speed as a
function of the simulation time
y
-0.5
-1
-1
x
Spatial distribution of the nodes
Mutual Interaction of Unicast and
Multicast Communication Session


Scenario

Single multicast VoIP session in the downlink direction

Increasing number of unicast VoIP session in the uplink direction

Effect of adding one unicast video feed in addition to the uplink VoIPs
Performance measurement



How does the performance of multicast session is degraded and vice
versa?
Metrics

Packet Deliver Ratio

PDR
PHY and MAC parameters, traffic emulation

VoIP: CBR G.711 with 10ms payload size (92 bytes/packet)

Video: CBR 30ms payload size (360Kbps)
Simulation Setup

Cell radiuses
1
2Mbps:
11Mbps:
Target SNR = 10dB
Cell Radius =240m
0.8
0.7
0.6
PER
Target SNR = 6dB
Cell Radius =300m
2Mbps
11Mbps
0.9
0.5
0.4
0.3
0.2
0.1
0
0
1
2
3
4
5
6
7
8
9
10
SNR (dB)

Probability of hidden
nodes


Carrier sensing range
(534m)
Hidden node probability 3%
0.03
0.025
0.02
0.015
0.01
0.005
0
0
20
40
60
80
100
120
140
Packet Delivery Ratio (PDR)
11Mbps
11Mbps - 1 video
0.92
0.9
0.88
0.86
0.84
2Mbps
2Mbps - 1 video
0.94
Multicast voice PDR
Multicast voice PDR
0.94
0.92
0.9
0.88
0.86
2
3
4
5
No. of unicast voice sessions
6
0.84
2
3
4
5
No. of unicast voice sessions
6
Delay
12
11Mbps
2 Mbps
11Mbps - 1 video
2 Mbps - 1 video
10
Unicast voice delay (ms)
Multicast voice delay (msec)
12
8
6
4
10
8
6
4
2
2
0
11Mbps
2 Mbps
11Mbps - 1 video
2 Mbps - 1 video
0
2
3
4
5
No. of unicast voice sessions
6
2
3
4
5
No. of unicast voice sessions
6
Final remarks


The degradation in performance can be concluded as

For multicast session it is the low PDR

For unicast sessions it is the increasing average delay
Future Work

It is expected that the degradation would be severe in the presence
of fading

It will be more appropriate to model the traffic instead of considering
the CBR type of traffic

DCF vs PCF
The End
Questions?
Thank You!