WBest: a Bandwidth Estimation Tool for IEEE 802.11 Wireless Networks

1
WBest a Bandwidth Estimation Tool for IEEE
802.11 Wireless Networks
Mingzhe Li, Mark Claypool, and Robert
Kinicki lmz, claypool, rek_at_cs.wpi.edu
Department of Computer Science, Worcester
Polytechnic Institute, Worcester MA, 01609 USA
Presented by Feng Li (lif_at_cs.wpi.edu)

• 33rd IEEE Conference on Local Computer Networks
  (LCN), Montreal, Quebec, Canada, October 16th,2008

2
Motivation

• Bandwidth estimation techniques focus on network
  capacity or available bandwidth.
• Most bandwidth estimation involved only wired
  networks.
• This paper presents a new Wireless Bandwidth
  estimation tool, WBest, designed for fast,
  non-intrusive, accurate estimation of available
  bandwidth over wireless LANs.

3
Challenges on Bandwidth Estimation

• Traditional approaches.
• (e.g. pathChirp v2.4.1 Ribeiro 2003, pathload
  v1.3.2 Jain 2003 etc.)
• Designed for precisely estimate the bandwidth in
  wired networks.
• Converge based on searching algorithms.
• Provide limited bandwidth information.
• Impacted by wireless networks.
• (e.g. shared media, retransmission, interference
  etc),
• Inaccurate results.
• Long estimation time.
• High intrusiveness.

4
Capacity Estimation with Packet Dispersion
Bottleneck router
L Packet size Ci Bottleneck capacity ?in Initi
al gap ?out Dispersed gap
5
Example Packet Dispersion with Wireless
Contention
Probing traffic
Contending traffic / Co-channel interference
6
Outline

• Motivation and Backgrounds
• WBest Algorithm
• Evaluation Experiments
• Result Analysis
• Conclusions

7
Terminology

• Effective Capacity (Ce )
• Maximum possible bandwidth that a link or
  end-to-end path can deliver.
• Available Bandwidth (A )
• Maximum unused bandwidth at a link or end-to-end
  path in a network.
• Typically, it is a time-varying metric.

8
Wireless Bandwidth Estimation Tool (WBest)

• Objective
• Fast, low intrusiveness, adequately accurate
  estimation of available bandwidth and variance of
  bandwidth in wireless networks.
• Two-step algorithm
• Packet pair technique to estimate effective
  capacity (Ce) of wireless network.
• Packet train technique to estimate mean and
  standard deviation of available bandwidth (A).

9
WBest Assumptions

• Assume last hop wireless network (hth hop) is
  bottleneck link with a single FCFS queue and

• Assume no significant changes in network
  conditions between two steps (estimating Ce and
  A).

10
Estimating Effective Capacity (Ce)

• Send n packet pairs to estimate Ce

• Ti dispersion time of ith packet pair
  (seconds),
• L packet size (bytes).

• Use median of n estimations to minimize impacts
  of crossing and contending traffic.

11
Estimating Available Bandwidth (A)

• A packet train of m packets is sent at effective
  capacity (Ce) to estimate available bandwidth (A).

• FCFS queuing at AP.

• R dispersion rate S crossing/contending
  traffic
• S reduced crossing/contending traffic

• Estimate contending and crossing traffic (S)
  using dispersion rate (R)

12
Estimating Available Bandwidth (A) (contd)

• Mean available bandwidth (A).

Fig 3 Estimating Available Bandwidth using
Average Dispersion Rate (R).
13
WBest Algorithm
1st Phase Calculating Ce
n 30
m 30
2nd Phase Calculating A
Error Correction
14
Outline

• Motivation and Background
• WBest Algorithm
• Evaluation Experiments
• Result Analysis
• Conclusions

15
Evaluation Setup
Client C

• Build testbed
• Open source drivers
• Wireless sniffer
• Various wireless conditions
• Traffic load
• Power saving mode
• Rate adaptation
• Implementation of WBest

• Compare with
• IGI/PTR v2.0 Hu 2003 (PGM/PRM)
• pathChirp v2.4.1 Ribeiro 2003 (PRM)
• pathload v1.3.2 Jain 2003
• (PRM)

16
Experiment Design

• 14 cases were designed to evaluate four bandwidth
  estimation tools under different network
  conditions.
• Each of 14 cases were repeated 30 times.
• All clients were placed with pre-selected
  locations with RSSI range between -38 and -42
  dBm.
• All experiments were run during summer break to
  eliminate effects from occasional wireless
  activities.

17
Result-Convergence Time vs. Error
18
Result-Intrusiveness vs. Error
19
Future Work

• Apply WBest to multimedia streaming applications
  to improve media performance and playout buffer
  optimization on wireless networks.
• Evaluate WBest performance under more complex
  wireless environments.
• Enhance WBest robustness during AP queue
  overflow.
• Develop new metric to replace Available Bandwidth
  (A) when TCP flows involved.

20
Conclusions

• Current bandwidth estimation tools are
  significantly impacted by wireless network
  conditions, such as contention or rate
  adaptations.
• Current tools are generally impractical for
  applications such as streaming multimedia that
  require fast, accurate and low intrusive
  bandwidth estimation.
• WBest consistently provides fast available
  bandwidth estimation, with generally more
  accurate estimates and lower intrusiveness under
  all conditions evaluated.

21
Question ?

• WBest with source code is available at
• http//perform.wpi.edu/downloads/wbest

22
Thank You!
WBest a Bandwidth Estimation Tool for IEEE
802.11 Wireless Networks
Mingzhe Li, Mark Claypool, and Robert
Kinicki lmz, claypool, rek_at_cs.wpi.edu
Department of Computer Science, Worcester
Polytechnic Institute, Worcester MA, 01609 USA
Presented by Feng Li (lif_at_cs.wpi.edu)
23
Reference

• Hu 2003 Ningning Hu and Peter Steenkiste,
  Evaluation and characterization of available
  bandwidth probing techniques, IEEE Journal on
  Selected Areas in Communications, vol. 21, no. 6,
  Aug. 2003.
• Ribeiro 2003 V. Ribeiro, R. Riedi, R. Baraniuk,
  J. Navratil, and L. Cottrell, pathchirp
  Efficient available bandwidth estimation for
  network paths, in PAM 03, La Jolla, CA, USA,
  Apr. 2003.
• Jain 2003 Manish Jain and Constantinos
  Dovrolis, End-to-end available bandwidth
  Measurement methodology, dynamics, and relation
  with tcp throughput, IEEE/ACM Transactions in
  Networking, , no. 295-308, Aug. 2003.

24
Analysis of Number of Packet Pairs
25
Analysis of Length of Packet Train