Scalable clustering of Internet paths by shared congestion

Min Sik Kim, Taekhyun Kim, Yong June Shin, Simon S. Lam, Edward J. Powers

Research output: Chapter in Book/Report/Conference proceedingConference contribution

9 Citations (Scopus)

Abstract

Internet paths sharing the same bottleneck can be identified using several shared congestion detection techniques. However, all of these techniques have been designed to detect shared congestion between a pair of paths. To cluster N paths by shared congestion, a straightforward approach of using pairwise tests would require O(N2) time complexity. In this paper, we present a scalable approach to cluster Internet paths based on DCW (Delay Correlation with Wavelet denoising) which does not require a common end point between paths. We present a function to map each path's measurement data into a point in a multidimensional space such that points are close to each other if and only if the corresponding paths share congestion. Because points in the space are indexed using a tree-like structure, the computational complexity of clustering N paths can be reduced to O(N log N). The indexing overhead can be further improved by reducing dimensionality of the space through wavelet transform. Computation cost is kept low by reusing for dimensionality reduction the same wavelet coefficients obtained in DCW. Our approach is evaluated by simulations and found to be effective for a large N. The tradeoff between dimensionality and clustering accuracy is shown empirically.

Original languageEnglish
Title of host publicationProceedings - INFOCOM 2006
Subtitle of host publication25th IEEE International Conference on Computer Communications
DOIs
Publication statusPublished - 2006 Dec 1
EventINFOCOM 2006: 25th IEEE International Conference on Computer Communications - Barcelona, Spain
Duration: 2006 Apr 232006 Apr 29

Publication series

NameProceedings - IEEE INFOCOM
ISSN (Print)0743-166X

Other

OtherINFOCOM 2006: 25th IEEE International Conference on Computer Communications
CountrySpain
CityBarcelona
Period06/4/2306/4/29

Fingerprint

Internet
Wavelet transforms
Computational complexity
Costs

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Electrical and Electronic Engineering

Cite this

Kim, M. S., Kim, T., Shin, Y. J., Lam, S. S., & Powers, E. J. (2006). Scalable clustering of Internet paths by shared congestion. In Proceedings - INFOCOM 2006: 25th IEEE International Conference on Computer Communications [4146794] (Proceedings - IEEE INFOCOM). https://doi.org/10.1109/INFOCOM.2006.141
Kim, Min Sik ; Kim, Taekhyun ; Shin, Yong June ; Lam, Simon S. ; Powers, Edward J. / Scalable clustering of Internet paths by shared congestion. Proceedings - INFOCOM 2006: 25th IEEE International Conference on Computer Communications. 2006. (Proceedings - IEEE INFOCOM).
@inproceedings{cb26d04342b74063a27a13a23b33b584,
title = "Scalable clustering of Internet paths by shared congestion",
abstract = "Internet paths sharing the same bottleneck can be identified using several shared congestion detection techniques. However, all of these techniques have been designed to detect shared congestion between a pair of paths. To cluster N paths by shared congestion, a straightforward approach of using pairwise tests would require O(N2) time complexity. In this paper, we present a scalable approach to cluster Internet paths based on DCW (Delay Correlation with Wavelet denoising) which does not require a common end point between paths. We present a function to map each path's measurement data into a point in a multidimensional space such that points are close to each other if and only if the corresponding paths share congestion. Because points in the space are indexed using a tree-like structure, the computational complexity of clustering N paths can be reduced to O(N log N). The indexing overhead can be further improved by reducing dimensionality of the space through wavelet transform. Computation cost is kept low by reusing for dimensionality reduction the same wavelet coefficients obtained in DCW. Our approach is evaluated by simulations and found to be effective for a large N. The tradeoff between dimensionality and clustering accuracy is shown empirically.",
author = "Kim, {Min Sik} and Taekhyun Kim and Shin, {Yong June} and Lam, {Simon S.} and Powers, {Edward J.}",
year = "2006",
month = "12",
day = "1",
doi = "10.1109/INFOCOM.2006.141",
language = "English",
isbn = "1424402212",
series = "Proceedings - IEEE INFOCOM",
booktitle = "Proceedings - INFOCOM 2006",

}

Kim, MS, Kim, T, Shin, YJ, Lam, SS & Powers, EJ 2006, Scalable clustering of Internet paths by shared congestion. in Proceedings - INFOCOM 2006: 25th IEEE International Conference on Computer Communications., 4146794, Proceedings - IEEE INFOCOM, INFOCOM 2006: 25th IEEE International Conference on Computer Communications, Barcelona, Spain, 06/4/23. https://doi.org/10.1109/INFOCOM.2006.141

Scalable clustering of Internet paths by shared congestion. / Kim, Min Sik; Kim, Taekhyun; Shin, Yong June; Lam, Simon S.; Powers, Edward J.

Proceedings - INFOCOM 2006: 25th IEEE International Conference on Computer Communications. 2006. 4146794 (Proceedings - IEEE INFOCOM).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Scalable clustering of Internet paths by shared congestion

AU - Kim, Min Sik

AU - Kim, Taekhyun

AU - Shin, Yong June

AU - Lam, Simon S.

AU - Powers, Edward J.

PY - 2006/12/1

Y1 - 2006/12/1

N2 - Internet paths sharing the same bottleneck can be identified using several shared congestion detection techniques. However, all of these techniques have been designed to detect shared congestion between a pair of paths. To cluster N paths by shared congestion, a straightforward approach of using pairwise tests would require O(N2) time complexity. In this paper, we present a scalable approach to cluster Internet paths based on DCW (Delay Correlation with Wavelet denoising) which does not require a common end point between paths. We present a function to map each path's measurement data into a point in a multidimensional space such that points are close to each other if and only if the corresponding paths share congestion. Because points in the space are indexed using a tree-like structure, the computational complexity of clustering N paths can be reduced to O(N log N). The indexing overhead can be further improved by reducing dimensionality of the space through wavelet transform. Computation cost is kept low by reusing for dimensionality reduction the same wavelet coefficients obtained in DCW. Our approach is evaluated by simulations and found to be effective for a large N. The tradeoff between dimensionality and clustering accuracy is shown empirically.

AB - Internet paths sharing the same bottleneck can be identified using several shared congestion detection techniques. However, all of these techniques have been designed to detect shared congestion between a pair of paths. To cluster N paths by shared congestion, a straightforward approach of using pairwise tests would require O(N2) time complexity. In this paper, we present a scalable approach to cluster Internet paths based on DCW (Delay Correlation with Wavelet denoising) which does not require a common end point between paths. We present a function to map each path's measurement data into a point in a multidimensional space such that points are close to each other if and only if the corresponding paths share congestion. Because points in the space are indexed using a tree-like structure, the computational complexity of clustering N paths can be reduced to O(N log N). The indexing overhead can be further improved by reducing dimensionality of the space through wavelet transform. Computation cost is kept low by reusing for dimensionality reduction the same wavelet coefficients obtained in DCW. Our approach is evaluated by simulations and found to be effective for a large N. The tradeoff between dimensionality and clustering accuracy is shown empirically.

UR - http://www.scopus.com/inward/record.url?scp=39049162846&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=39049162846&partnerID=8YFLogxK

U2 - 10.1109/INFOCOM.2006.141

DO - 10.1109/INFOCOM.2006.141

M3 - Conference contribution

AN - SCOPUS:39049162846

SN - 1424402212

SN - 9781424402212

T3 - Proceedings - IEEE INFOCOM

BT - Proceedings - INFOCOM 2006

ER -

Kim MS, Kim T, Shin YJ, Lam SS, Powers EJ. Scalable clustering of Internet paths by shared congestion. In Proceedings - INFOCOM 2006: 25th IEEE International Conference on Computer Communications. 2006. 4146794. (Proceedings - IEEE INFOCOM). https://doi.org/10.1109/INFOCOM.2006.141