Designing An 802.11g Ad-hoc Network for Multimedia Communication

1
Designing An 802.11g Ad-hoc Network for
Multimedia Communication

• Chung-Wei Lee Jonathan C.L. Liu
• Presented By Mahendra Kumar

2
Layout

• Introduction
• Wireless network problems
• Performance of Ad-hoc network in different
  environment- Result and Analysis
• Proposed scheme for routing decision- Max
  throughput
• Simulation results
• Conclusion
• Future work

3
Introduction

• Ad-hoc mode 54 Mbps theoretical limit
• Route optimization to support multimedia
  communication.
• Has a mixed environment depending upon the path.
• Is Min-Hop count algorithm okay?
• Focus should be in maximizing the overall
  throughput from source to destination.

4
Wireless network problems

• Ad-hoc operates on 802.11g wireless mode.
• Has to deal with channel errors due to
  propagation loss, reflection, interference.
• Indoor and outdoor environment have different
  benefits and looses.
• Indoor- signal has more way to propagate due to
  reflection. But loss due to scattering and
  obstacles.
• Outdoor- signal suffers due to natural obstacles
  like temperature and humidity. Interference from
  other sources.

5
Wireless network problems(Contd)

• Hard to predict how the wireless network will
  perform under these complex environmental
  parameters.
• Routing algorithm for ad-hoc should be adaptive
  for these different environment.
• Conventionally, the network bandwidth should be
  higher with decrease in distance.
• Experiment and analysis shows that performance
  is worst within 5 meters distance in indoor.
• Design a algorithm which takes into all these
  problems.

6
Performance of Ad-hoc network

• Two identical laptops with Pentium 4 processor,
  512 memory, linksys 802.11g, 2.4 GHz wireless
  card set in ad-hoc mode.
• Custom benchmarking tool built on top of the
  existing TCP/UDP protocol stack.
• Three different Environment
• Indoor without obstruction
• Outdoor without obstruction
• Penetrating wall

7
Performance of ad-hoc network

• Indoor without obstruction
• Department basement
• Three distances
• Within 5m, 5m - 10m, 10m 20m
• Outdoor without obstruction
• Parking lot
• Distances of 5, 10, 15, 20 and 25 meters
• Penetrating wall
• One indoor laptop, other laptop outside such that
  there is a wall between them.
• Change the distance between them from 5, 10, 15
  and 20.

8
Indoor TCP without obstruction

• Performance increase with distance. Why?
• larger messages cause smaller software overhead
• 19 throughput improvement by increasing pkt size
  from 32KB to 4 MB

9
Indoor UDP without obstruction

• Performs better than TCP. Why?
• Less overhead than TCP

10
Outdoor TCP without obstruction

• Distance trend not similar to the Indoor
  behavior. Why?
• Less multi-path interference.

11
Outdoor UDP without obstruction

• Similar trend as outdoor TCP but improved
  throughput.

12
Penetrating Wall-TCP

• Throughput decrease with increase in distance.

13
Proposed Scheme

• Throughput-distance is the major concern in QoS
  routing in ad-hoc.
• Hence routing should be done based on maximum
  throughput and not on min-hop count.
• Majority of existing schemes do not take distance
  and throughput into account.

14
Generating simulating nodes

• Random topology algorithm

15
Max Throughput algorithm
16
Simulation design parameters

• Generating n random nodes in an area of 300 by
  300 meters using random topology algorithm.
• Calculate the end to end bandwidth for all
  possible node pairs.
• Calculate average bandwidth over all paths. i.e.
  n(n-1)/2 paths and 100 topologies.

17
Comparing Average Path bandwidth
18
Path bandwidth distribution
19
Conclusion

• Proposed scheme of max throughput performs better
  than min hop count.
• Better suitable for multimedia based traffic.
• Higher number of nodes gives more average
  bandwidth between source and destination in max
  throughput.

20
Future Work

• Simulation on different operating systems.
• Routing nodes are mobile in ad-hoc.
• Some nodes can move to appropriate position to
  increase the throughput.
• On the contrary, some may lose their connection.
• Find a novel scheme which can optimize the
  position of nodes to guarantee max throughput.
• Exploring a routing scheme based on combination
  of Min-hop and Max-throughput and min-distance
  algorithm.

21
Questions?