Estimating Link Capacity in High Speed Networks

1
Estimating Link Capacity in High Speed Networks

• Ling-Jyh Chen1, Tony Sun2, Li Lao2, Guang Yang2,
  M.Y. Sanadidi2, Mario Gerla2
• 1Institute of Information Science, Academia
  Sinica
• 2Dept. of Computer Science, University of
  California at Los Angeles

2
Definition

• Capacity maximum IP-layer throughput that a flow
  can get, without any cross traffic.
• Available Bandwidth maximum IP-layer throughput
  that a flow can get, given (stationary) cross
  traffic.

3
Previous Work on Capacity Estimation

• Per-hop based
• pathchar use different packet sizes to probe the
  per-hop link capacity
• clink, pchar variants of pathchar
• Nettimer use packet tailgating technique
• End-to-end based
• Pathrate, Sprobe, CapProbe
• For specialized networks AsymProbe, ALBP, AdHoc
  Probe
• How about high speed networks?

4
Estimating High Speed Links

• High speed links are becoming popular (e.g. GB
  links, DVB links, and UWB links)
• However, capacity estimation on high speed links
  remains a challenge (e.g., probing pksize and
  system time resolution are limited)
• And, an effective estimation tool for high speed
  links is still desired

5
Our Contribution

• We propose an end-to-end capacity estimation
  technique for high speed links, called PBProbe.
• PBProbe is based on CapProbe
• One-way method
• UDP based
• packet bulk based
• simple, fast, and accurate

6
Packet Pair Dispersion
Capacity (Packet Size) / (Dispersion)
7
Issues Compression and Expansion

• Queueing delay on the first packet gt
  compression
• Queueing delay on the second packet gt expansion

8
CapProbe (Rohit et al, SIGCOMM04)

• Key insight a packet pair that gets through with
  zero queueing delay yields the exact estimate.
• CapProbe uses Minimum Delay Sum filter.

9
Proposed Approach PBProbe

• Have two phases for both forward and backward
  link estimation
• Use packet bulk (instead of packet pair) of
  length k in each probing
• Adapt k to enlarge the dispersion between the
  first and last packet, and thus overcome the
  timer resolution problem
• Tradeoff BW consumption and estimation speed by U
  parameter

10
Proposed Approach PBProbe
11
Proposed Approach PBProbe

• k is depended on the estimated link capacity and
  the supported timer resolution.
• n is set to 200.
• Dthresh is set to 1ms.
• U is set to 0.002.

12
Analysis

• Poisson cross traffic (arrival and service rates
  are ? and µ), service time is t
• Prob. of obtaining a good sample
• Expected number of samples required for obtaining
  a good sample

13
Analysis
14
Evaluation

• NISTNet emulation
• High speed Internet experiments
• Comparison of PBProbe and Pathrate

15
Evaluation 1 NISTNet emulation

• No cross traffic

16
Evaluation 2 Internet experiments

• 5 hosts NTNU, UCLA, CalTech, GaTech, PSC(n
  200, k 100, 20 runs)

17
Evaluation 3 PBProbe vs Pathrate
18
Summary

• We propose an end-to-end capacity estimation
  technique, called PBProbe, for high speed links.
• We evaluate PBProbe by analysis, emulation and
  Internet experiments.
• We show that PBProbe can correctly and rapidly
  estimate bottleneck capacity in almost all test
  cases.

19
Acknowledgements

• This work is co-sponsored by the National Science
  Council and the National Science Foundation under
  grant numbers NSC-94-2218-E-001-002 and
  CNS-0435515.
• We are grateful to the following people for their
  help in carrying out PBProbe measurements Sanjay
  Hegde (CalTech), Che-Chih Liu (NTNU), Cesar A. C.
  Marcondes (UCLA), and Anders Persson (UCLA).

20
Thanks!
CapProbe http//nrl.cs.ucla.edu/CapProbe/