2017
Karyotis, V.; Stai, E.
Hyperbolic big data analytics for dynamic network management and optimization Book
CRC Press, 2017, ISBN: 9781498784870; 9781498784863, (cited By 1).
Abstract | Links | BibTeX | Tags: Big data; Environmental management; Machinery; Network layers; Network management, Dynamic network management; Efficient managements; Embedding of graphs; Multi-layer network; Network structures; Networking environment; Operational environments; Structural feature, Information management
@book{Karyotis2017177,
title = {Hyperbolic big data analytics for dynamic network management and optimization},
author = {V. Karyotis and E. Stai},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85052714130&doi=10.1201%2fb21278&partnerID=40&md5=538cccccfa15d20aa1230ee5bb1891df},
doi = {10.1201/b21278},
isbn = {9781498784870; 9781498784863},
year = {2017},
date = {2017-01-01},
journal = {Big Data and Computational Intelligence in Networking},
pages = {177-207},
publisher = {CRC Press},
abstract = {178Massive numbers of devices, growing user populations and voluminous amounts of produced/exchanged information are expected in the complex cyber-physical networks of the future. The new scales of operation give rise to a big network data era, where multi-layer networks form and users become content producers/consumers (prosumers). The challenges associated with the forthcoming networking environments require radical rethinking of current network analysis, management and operation practices. This chapter will focus on this effort, and more specifically on reinventing the machinery for computing key network metrics that allow improving network management/operation, while also developing more efficient overlay applications within demanding operational environments. Special attention is given to the impact of network evolution on the computation of such metrics, an aspect relatively neglected until recently. In order to efficiently compute key metrics associated with social or structural features of the network and track them when the infrastructure evolves, a big data analytics methodology, denoted as Hyperbolic Data Analytics (HDA), is applied. HDA is based on the embedding of graphs in the hyperbolic space, leading to more efficient management/operation, typically by exploiting hidden network structure. HDA is a characteristic example of developing computational intelligence over big network data and exploiting them for improving/optimizing both infrastructures and applications/services. HDA will provide the means for computing efficiently network analysis metrics, the evolution of which indicates the evolution of the corresponding network’s structure. © 2018 by Taylor and Francis Group, LLC.},
note = {cited By 1},
keywords = {Big data; Environmental management; Machinery; Network layers; Network management, Dynamic network management; Efficient managements; Embedding of graphs; Multi-layer network; Network structures; Networking environment; Operational environments; Structural feature, Information management},
pubstate = {published},
tppubtype = {book}
}
178Massive numbers of devices, growing user populations and voluminous amounts of produced/exchanged information are expected in the complex cyber-physical networks of the future. The new scales of operation give rise to a big network data era, where multi-layer networks form and users become content producers/consumers (prosumers). The challenges associated with the forthcoming networking environments require radical rethinking of current network analysis, management and operation practices. This chapter will focus on this effort, and more specifically on reinventing the machinery for computing key network metrics that allow improving network management/operation, while also developing more efficient overlay applications within demanding operational environments. Special attention is given to the impact of network evolution on the computation of such metrics, an aspect relatively neglected until recently. In order to efficiently compute key metrics associated with social or structural features of the network and track them when the infrastructure evolves, a big data analytics methodology, denoted as Hyperbolic Data Analytics (HDA), is applied. HDA is based on the embedding of graphs in the hyperbolic space, leading to more efficient management/operation, typically by exploiting hidden network structure. HDA is a characteristic example of developing computational intelligence over big network data and exploiting them for improving/optimizing both infrastructures and applications/services. HDA will provide the means for computing efficiently network analysis metrics, the evolution of which indicates the evolution of the corresponding network’s structure. © 2018 by Taylor and Francis Group, LLC.