Network topology discovery using finite mixture models

Author

Shih, Meng-Fu ; Hero, Alfred O.

Author_Institution

Dept. of EECS, Michigan Univ., Ann Arbor, MI, USA

Volume

2

fYear

2004

fDate

17-21 May 2004

Abstract

We propose a network topology estimation strategy using unicast end-to-end packet pair delay measurements that is based on mixture models for the delay covariances. An unsupervised learning algorithm is applied to estimate the number of mixture components and delay covariances. The leaf pairs are clustered by a MAP criterion and passed to a hierarchical topology construction algorithm to rebuild the tree. Results from an ns simulation show that our algorithm can identify a network tree with 8 leaf nodes.

Keywords

Internet; covariance analysis; delays; maximum likelihood estimation; network topology; telecommunication computing; trees (mathematics); unsupervised learning; Internet; MAP criterion; delay covariance; end-to-end delay; finite mixture models; hierarchical topology construction; leaf pairs; network topology estimation strategy; network tree leaf nodes; packet pair delay; unicast; unsupervised learning algorithm; Clustering algorithms; Delay estimation; Inference algorithms; Internet; Network topology; Routing; Telecommunication traffic; Tomography; Unicast; Unsupervised learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 2004. Proceedings. (ICASSP '04). IEEE International Conference on

ISSN

1520-6149

Print_ISBN

0-7803-8484-9

Type

conf

DOI

10.1109/ICASSP.2004.1326287

Filename

1326287