Application of graph theory to clustering in delay space

Author

Zhanikeev, Marat ; Tanaka, Yoshiaki

Author_Institution

Sch. of Int. Liberal Studies, Waseda Univ., Tokyo, Japan

fYear

2010

fDate

15-18 June 2010

Firstpage

1

Lastpage

6

Abstract

In recent years, many methods have been proposed for analysis of end-to-end delay space generated when end-to-end measurements are conducted among a large number of globally distributed nodes without the knowledge of underlying network topology. Most analysis in this area concludes that coordinate space generated from end-to-end delay has from 4 to 7 dimensions, based on frequency components found by PCA or similar techniques. This paper takes a traditional graph theory approach to clustering and proves that even with small number of end-to-end measurement data it is possible to detect clusters.

Keywords

delays; graph theory; principal component analysis; telecommunication network topology; PCA; clustering; end-to-end delay space; end-to-end measurements; graph theory; network topology; principle component analysis; Clustering algorithms; Delay; Extraterrestrial measurements; Graph theory; IP networks; Network topology; Principal component analysis; Social network services; Telecommunication network topology; Telecommunication traffic; clustering; end-to-end delay space; graph theory; matrix transform; network embedding;

fLanguage

English

Publisher

ieee

Conference_Titel

Information and Telecommunication Technologies (APSITT), 2010 8th Asia-Pacific Symposium on

Conference_Location

Kuching

Print_ISBN

978-1-4244-6413-5

Electronic_ISBN

978-4-88552-244-4

Type

conf

Filename

5532047