particle algorithms, prediction, user-profiles, VoIP
Abstract :
[en] An important problem one needs to deal with in
a Voice over IP system is server overload. One way for pre-
venting such problems is to rely on prediction techniques for
the incoming traffic, namely as to proactively scale the avail-
able resources. Anticipating the computational load induced
on processors by incoming requests can be used to optimize
load distribution and resource allocation. In this study, the
authors look at how the user profiles, peak hours or call pat-
terns are shaped for a real system and, in a second step, at
constructing a model that is capable of predicting trends.
Disciplines :
Computer science
Author, co-author :
SIMIONOVICI, Ana-Maria ; University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC)
Tantar, Alexandru; University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC)
BOUVRY, Pascal ; University of Luxembourg > Faculty of Science, Technology and Communication (FSTC) > Computer Science and Communications Research Unit (CSC)
Didelot, Loic; MIXvoip S.a, Luxembourg
External co-authors :
no
Language :
English
Title :
Predictive Modeling in a VoIP System
Publication date :
September 2013
Journal title :
Journal of Telecommunications and Information Technology
ISSN :
1509-4553
eISSN :
1899-8852
Publisher :
National Institute of Telecommunications, Warsaw, Poland