Evaluating pathrate and pathload with realistic cross-traffic Ravi - - PowerPoint PPT Presentation

▶

Nov 10, 2023 246 likes •447 views

Evaluating pathrate and pathload with realistic cross-traffic Ravi Prasad Manish Jain Constantinos Dovrolis (ravi, jain, dovrolis@cc.gatech.edu) College of Computing Georgia Institute of Technology Bandwidth Estimation Workshop 2003

SLIDE 1

Bandwidth Estimation Workshop 2003

Evaluating pathrate and pathload with realistic cross-traffic

Ravi Prasad Manish Jain Constantinos Dovrolis (ravi, jain, dovrolis@cc.gatech.edu) College of Computing Georgia Institute of Technology

SLIDE 2

Bandwidth Estimation Workshop 2003

Background

!Pathrate

!Estimates path capacity !Based on packet pair/train dispersion

!Packet pair estimates: Set of possible capacity modes !Packet train estimates: ADR=Lower bound on capacity

!Capacity = (Strongest and narrowest mode > ADR)

!Pathload

!Estimates path available bandwidth (avail-bw) !Based on one-way delay trend of periodic streams !Reports a range of avail-bw

!Corresponds to variation, measured in stream duration

!http://www.pathrate.org

SLIDE 3

Bandwidth Estimation Workshop 2003

Motivation

!Recent studies pointed towards poor accuracy

f these tools

!http://www.caida.org/outreach/presentations/200 3/bwest0308/doereview.pdf !A measurement study of available bandwidth estimation tools. Strauss et. al. IMC 2003

!Our objective: re-evaluate accuracy of both tools

!Wide range of cross-traffic load !Realistic cross-traffic !Completely monitored testbed (no guessing!)

SLIDE 4

Bandwidth Estimation Workshop 2003

Outline

!Describe test methodology

!Testbed !Cross-traffic type

!Show accuracy results

!100Mbps path !1Gbps path !With Iperf cross-traffic

!Explaining inaccuracies with Iperf cross- traffic !Conclusions

SLIDE 5

Bandwidth Estimation Workshop 2003

Testing methodology

!Used local testbed

!Complete knowledge of path properties

!Capacity !Available bandwidth

!Complete control of cross-traffic

! Rate ! Type (TCP vs UDP vs trace-driven)

SLIDE 6

Bandwidth Estimation Workshop 2003

Testbed

! Narrow link capacity C = 100Mbps or 1Gbps

1 Gbps 1 Gbps C Cross−traffic Measurement Traffic

Workgroup Switch

Catalyst

Workgroup Switch

Catalyst

CiscoSystems CiscoSystems

SLIDE 7

Bandwidth Estimation Workshop 2003

Cross traffic

! Trace-driven cross-traffic generation:

! NLANR traces

!OC-3, OC-12, OC-48

! Trace information at the end of the talk

! Packet size distribution

!Unmodified

! Packet interarrivals

!Either, scaled to achieve desired cross-traffic throughput !Or, unmodified

! Iperf-based cross-traffic

! Single TCP stream ! UDP stream

SLIDE 8

Bandwidth Estimation Workshop 2003

Results

SLIDE 9

Bandwidth Estimation Workshop 2003

FastEthernet: Traces with scaled interarrivals

20 40 60 80 100

Cross-traffic utilization (Mbps)

20 40 60 80 100

Estimated bandwidth (Mbps)

Pathrate Pathload Capacity Avail bw

SLIDE 10

Bandwidth Estimation Workshop 2003

FastEthernet: Traces with unmodified interarrivals

20 40 60 80 100

Cross-traffic utilization (Mbps)

20 40 60 80 100

Estimated bandwidth (Mbps)

Pathrate Pathload Capacity Avail bw

SLIDE 11

Bandwidth Estimation Workshop 2003

Gigabit path: Traces with scaled interarrivals

200 400 600 800 1000

Cross-traffic utilization (Mbps)

200 400 600 800 1000

Estimated bandwidth (Mbps)

Pathrate Pathload Capacity Avail bw

SLIDE 12

Bandwidth Estimation Workshop 2003

Unrealistic cross-traffic

! Single stream TCP

! Entire window appears as burst at beginning of RTT ! Minimum averaging interval: RTT

TR TW

! UDP periodic stream

! Packet size: L ! Rate: R ! Dispersion: L/R ! Utilization ρ = R/C

L/C L/R

SLIDE 13

Bandwidth Estimation Workshop 2003

Gigabit Path: Iperf Periodic UDP

200 400 600 800 1000

Cross-traffic utilization (Mbps)

200 400 600 800 1000

Estimated bandwidth (Mbps)

Pathrate Pathload Capacity Avail bw

SLIDE 14

Bandwidth Estimation Workshop 2003 L/C L/R TR TW TO

! Seeks some “off” time periods

f duration larger than L/C

! L: Probe size

Pathrate under unrealistic traffic

! UDP periodic traffic ! If ρ < 0.5 then TO > L/C ! Else, underestimation ! TCP traffic ! Off period TO =TR - TW- L/C

!Correct capacity estimate when TO > L/C

SLIDE 15

Bandwidth Estimation Workshop 2003

Gigabit path: Iperf single stream TCP

200 400 600 800 1000

Cross-traffic utilization (Mbps)

200 400 600 800 1000

Estimated bandwidth (Mbps)

Pathrate Pathload Capacity Avail bw

SLIDE 16

Bandwidth Estimation Workshop 2003 L/C L/R TR TW

Pathload under unrealistic traffic

! Samples avail-bw in stream duration (TS)

! UDP periodic traffic

! Avail-bw averaging period L/R ! TS = 100 x L/C > L/R ! Correct avail-bw range estimate

TS TS TS TS

! TCP traffic

! Avail-bw averaging period TR ! TS << TR results in wide Avail-bw range estimate

SLIDE 17

Bandwidth Estimation Workshop 2003

Conclusions

!Type of cross-traffic is important for bandwidth estimation tools !Pathrate and pathload perform well with realistic cross-traffic !Simulated traffic does not capture:

!Packet size distribution !Interarrival distribution !Correlation structure

SLIDE 18

Bandwidth Estimation Workshop 2003

Trace identifiers

! OC3 : MEM-1070464136-1, COS-1070488076-1, BWY-1063315231-1, COS-1049166362-1 ! OC12: MRA-1060885637-1 ! OC48: IPLS-CLEV-20020814-093000-1 ! We greatly appreciate the availability of traces from NLANR PMA project. The NLANR PMA project is supported by National Science Foundation Cooperative agreement nos. ANI-0129677 (2002) and ANI-9807479 (1998).