Reference architecture for social
networks graph analysis 1Laboratory
of Computer Security Problems,
St. Petersburg Institute for Informatics and Automation (SPIIRAS), 199178, St.
Petersburg, Russia {kolomeec, chechulin, kotenko}@comsec.spb.ru 2LAAS
CNRS, Université de
Toulouse, 31031, Toulouse, France {abenachour, elbaz}@laas.fr 3Université
de Toulouse – IRIT, 31400,
Toulouse, France martin.strecker@irit.fr 4ITMO
University, 197101,
Saint-Petersburg, Russia 5University of Sciences and Technology Houari Boumediene, BP 32 El-Alia, 16111, Algiers, Algeria Abstract When analyzing social networks, graph data
structures are often used. Such graphs may have a complex structure that
makes their operational analysis difficult or even impossible. This paper
discusses the key problems that researchers face in the field of processing big
graphs in that particular area. The paper proposes a reference architecture
for storage, analysis and visualization of social network graphs, as well as
a big graph process “pipeline”. Based on this pipeline it is possible to
develop a tool that will be able to filter, aggregate and process in parallel
big graphs of social networks, and at the same time take into account its
structure. The paper includes the implementation of that pipeline using the OrientDB graph database for storage, parallel processing
for graph measures calculation and visualization of big graphs using the D3
library. The paper also includes the conducted experiments based on the
calculation of betweenness centrality of some
graphs collected from the VKontakte social net. Keywords: Social
networks, graph analysis, big data, parallel computing, GPU, graph databases,
+: Corresponding author: Igor Kotenko
|
Journal of Wireless
Mobile Networks, Ubiquitous Computing, and Dependable Applications (JoWUA)
Vol.
10, No. 4, pp.109-125, December 2019 [pdf]
Received:
October 3, 2019; Accepted: December 12, 2019; Published: December 31, 2019
DOI: 10.22667/JOWUA.2019.12.31.109