2016
Karyotis, V.; Stai, E.; Papavassiliou, S.
On the evolution of complex network topology under network churn Journal Article
In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9674, pp. 227-240, 2016, ISSN: 03029743, (cited By 1; Conference of 14th IFIP WG 6.2 International Conference on Wired/Wireless Internet Communications, WWIC 2016 ; Conference Date: 25 May 2016 Through 27 May 2016; Conference Code:176559).
Abstract | Links | BibTeX | Tags: Analysis frameworks; Clustering coefficient; Edge churn; Future internet; Heterogeneous access; Multi-dimensional vectors; Network evolution; Network properties, Complex networks, Internet; Topology; Vector spaces
@article{Karyotis2016227,
title = {On the evolution of complex network topology under network churn},
author = {V. Karyotis and E. Stai and S. Papavassiliou},
editor = {Mamatas L. Matta I. Papadimitriou P.},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-84978959144&doi=10.1007%2f978-3-319-33936-8_18&partnerID=40&md5=d66eefa06bcaa59521adbf8ba126608e},
doi = {10.1007/978-3-319-33936-8_18},
issn = {03029743},
year = {2016},
date = {2016-01-01},
journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
volume = {9674},
pages = {227-240},
publisher = {Springer Verlag},
abstract = {The future Internet is becoming more diverse, incorporating heterogeneous access networks. The latter are characterized by numerous devices that join/leave the network dynamically, creating intense churn patterns. New approaches to analyze and quantify churn-induced network evolution are required. In this paper, we address such need by introducing a new analysis framework that maps network evolution into trajectories in multi-dimensional vector spaces. Each network instance is characterized by a feature vector, indicating network properties of interest. To demonstrate the potentials of this approach, we exemplify and study the effect of edge churn on various complex topologies, frequently emerging in various communications environments. We investigate via simulation the impact of network evolution, by quantifying its effect on key network analysis metrics, such as the clustering coefficient and the plethora of centrality metrics, employed at large for analyzing topologies and designing applications. The proposed framework aspires to establish more holistic and efficient complex network control. © IFIP International Federation for Information Processing 2016.},
note = {cited By 1; Conference of 14th IFIP WG 6.2 International Conference on Wired/Wireless Internet Communications, WWIC 2016 ; Conference Date: 25 May 2016 Through 27 May 2016; Conference Code:176559},
keywords = {Analysis frameworks; Clustering coefficient; Edge churn; Future internet; Heterogeneous access; Multi-dimensional vectors; Network evolution; Network properties, Complex networks, Internet; Topology; Vector spaces},
pubstate = {published},
tppubtype = {article}
}
The future Internet is becoming more diverse, incorporating heterogeneous access networks. The latter are characterized by numerous devices that join/leave the network dynamically, creating intense churn patterns. New approaches to analyze and quantify churn-induced network evolution are required. In this paper, we address such need by introducing a new analysis framework that maps network evolution into trajectories in multi-dimensional vector spaces. Each network instance is characterized by a feature vector, indicating network properties of interest. To demonstrate the potentials of this approach, we exemplify and study the effect of edge churn on various complex topologies, frequently emerging in various communications environments. We investigate via simulation the impact of network evolution, by quantifying its effect on key network analysis metrics, such as the clustering coefficient and the plethora of centrality metrics, employed at large for analyzing topologies and designing applications. The proposed framework aspires to establish more holistic and efficient complex network control. © IFIP International Federation for Information Processing 2016.