CONTENTION WINDOW CONTROL IN WRCP PROTOCOL FOR IEEE802'11 MULTIHOP AD HOC NETWORKS

1
CONTENTION WINDOW CONTROL IN WRCP PROTOCOL FOR
IEEE 802.11 MULTI-HOP AD HOC NETWORKS

• Mariusz BEDNARCZYK
• Marek AMANOWICZ
• Military University of Technology
• Telecommunication Institute
• Warsaw, Poland

2
Outline

• Introduction
• Description of the Wireless Relay Control
  Protocol
• Motivation
• The proposed tuning of CW
• Performance results
• Conclusion

3
Goals and limitations

• IEEE 802.11 MAC layer
• range limited to 1-hop
• routing as a great challange
• QoS, energy conservation, security

• move Routing and Forwarding functions to the
  MAC layer
• routing protocol taking into account
  multidimensional metrics
• fast data transferring via intermediate nodes
• low protocol complexity

4
The WRCP protocol description
B
A
C
E
D
If the path is not updated during 5s interval,
the node removes unused path
5
Future prospects

• mission based networks where units
• are expected to cooperate
• operate in a fixed area for
• a predetermined period
• move with less randomness

6
A closer look..
7
Problem

• short backoff mechanism for the relaying station
• it causes that in the network appear selfish
  users
• they grab the idle channel faster than others
• it forces legacy users for deferring their own
  transmission

Binary Exponential Backoff Algorithm
8
Solution

• Gentle Increase/Decrease Backoff Algorithm
• after an unsuccessful transmission, reasonably
  increase of CW rather than doubling it each time
• after a successful transmission, a sequential
  decrease of CW rather than resetting it to an
  initial value

9
Performance evaluation
10
Simulation Setup

• Omnetpp ver. 3.3 simulation tool
• WaveLAN IEEE 802.11 MAC
• nodes uniformly distributed on the circle and one
  node at the centre
• ring radius 200 m, max transmission distance 250
  m
• common power level for all nodes
• channel bandwidth 2 Mbps
• CWmin 7 for WRCP and CWmin 31 for legacy DCF
• size of data payload 1024 B
• RTS/CTS not used

11
Per-Flow Throughput vs. Offered Load for 2-hops
transmissions
12
Packet Delivery Ratio vs. Offered Load for 2-hops
transmissions
13
Average Packet Delay vs. Offered Load for 2-hops
transmissions
14
Average Packet Delay vs. Offered Load for 2-hops
transmissions in heavy traffic environment
15
Average Packet Delay vs. Offered Load for
well-behaved node in the vicinity of the selfish
node
16
Average Packet Delay vs. Offered Load with 3
selfish nodes next to each other
17
Packet Delivery Ratio vs. Offered Load with 3
selfish nodes next to each other
18
Conclusions
19
Analysis

• GIDB algorithm is simple and easy to adapt
• GIDB algorithm outperforms the Mac 802.11 DCF in
  terms of packet delay, throughput and packet loss
  rate
• high packet delivery success rate
• delay performance is acceptable for time critical
  data delivery

20
Open Issues

• QoS support
• service differentiation is not considered
• Mobility
• all the proposed work refer essentially to a
  static scenario
• IEEE 802.11 driver card modification

21
Thanks