Experimental and Analytical Evaluation of Available Bandwidth Estimation Tools
Cesar D. Guerrero and Miguel A. Labrador
Department of Computer Science & Engineering
University of South Florida

• Motivation
• Problem
• Testbed Description
• Analytical Model
• Target Tools
• Performance Evaluation
• Conclusions
Outline

Motivation
Why to evaluate available bandwidth tools?
• Available bandwidth  to improve network applications performance.
• Applications  different time, accuracy, and overhead from estimators.
• Evaluation  determine whether a tool is suitable for an application.

Problem
What issues do we want to solve?

Topology

Link capacities

Packet loss rate

Delay
• Evaluate tools over the same variable network conditions
• Analytical model to have a theoretical value to compare with

Testbed Description
Architecture
Low cost
Link A
192.168.3.0/24
Link B
192.168.2.0/24
Link C
192.168.1.1/24
Link D
192.168.0.0/24
192.168.4.0/24
Client
Probing packets Cross Traffic Server
• Client and server hosting bandwidth estimation tools
• Intermediate nodes hosting a packet shaper and a traffic generator
• Phython applications running in all the machines to automatically perform experiments.
• Internet connected

Analytical Model
Jackson Network
Probing packets Cross Traffic
λ
8
8
λ
1
1
λ
2
2
λ
3
3
λ
4
4
λ
5
5
λ
6
6 λ
7
7
Client
λ
0
γ
1
γ
3
γ
5
γ
7
• Eight M/M/1 queues model input and output packet flows.
• The Jackson model gives the average arrival rate to a node
λ j
= γ j
+
Σ λ i
θ ij
• The available bandwidth is the minimum non utilized capacity of the queues associated to the links:
A = min i= 1,3,5,7
( A i
) = min i= 1,3,5,7
(1-ρ i
)
Server

Target Tools
Estimation Approaches
Probe Rate Model
• • Pathload.
• TOPP
• Pathchirp
• PTR
Probe Gap Model
• Delphy

Target Tools
Pathload grey region
Figure copied from the paper “
Pathload: A Measurement Tool for
End-to-end Available Bandwidth
” by M. Jain and C. Dovrolis
• Fleet of probing streams are sent to fill the available bandwidth.
• The one-way delay increases when the rate of the probing traffic is higher that the available bandwidth.
• In the gray region , the tool reports the available bandwidth

Target Tools
IGI turning point
• Estimates the cross traffic as a function of the amount of traffic inserted between a packet pair.
• Available bandwidth is given by the average rate of the packet train when the initial packet gap is equal to the output gap.
Figure copied from the paper “ Evaluation and
Characterization of Available Bandwidth Probing
Techniques
” by N. Hu and P. Steenkiste

Target Tools
Spruce
• Probing packets are sent with an intra-pair gap ( Δ in
) equal to the narrow link transmission time of a 1500B packet (to guarantee that the pair will be in the queue at the same time)
• Cross traffic is measured using the dispersion of the probing packets
( Δ out
) calculated at the receiver.
• It requires a previous calculation of the tight link capacity (
C )
A

C


 1

 out

 in
 in



Performance Evaluation
Experiments
• Metrics: accuracy, time, overhead
• 28 network scenarios : link capacities from 1 to 10 Mbps and from 10 to
100 Mbps
• Each scenario with four cross traffic loads : 0%, 25%, 50%, and 75% of the capacity
• Every estimation was performed 35 times
• Accuracy plots have a 95% confidence interval
11760 experiments

Performance Evaluation
Accuracy with 75% of the capacity as cross traffic
Estimated available bandwidth / total bandwidth (capacity)
Pathload IGI Spruce

Performance Evaluation
Relative Error
β  experiment al
A
 analytical analytical
A
A
Pathload IGI Spruce

Performance Evaluation
Convergence Time
Pathload IGI Spruce

Performance Evaluation
Overhead
Probing traffic / total bandwidth (capacity) in the tight link
Pathload IGI Spruce

Conclusions
• Main contributions:
 Low cost and flexible testbed to evaluate estimation tools in a controlled network.

Analytical model to fairly compare the tools accuracy with a theoretical value.
• Regarding to the tools evaluation:

Pathload is the most accurate tool but the slowest to converge

IGI is the fastest tool but the least accurate

Spruce is the least intrusive tool with intermediate accuracy and convergence time.

Experimental and Analytical Evaluation of Available Bandwidth Estimation Tools
Cesar D. Guerrero cguerrer@cse.usf.edu
Miguel A. Labrador labrador@cse.usf.edu
Department of Computer Science & Engineering
University of South Florida