Traffic modelling for Big Data backed telecom cloud

The objective of this project is to provide traffic models based on new services characteristics. Specifically, we focus on modelling the traffic between origin-destination node pairs (also known as OD pairs) in a telecom network. Two use cases are distinguished: i) traffic generation in the context...

Descripción completa

Detalles Bibliográficos
Autor: Via Baraldés, Anna
Tipo de recurso: tesis de maestría
Fecha de publicación:2016
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/89946
Acceso en línea:https://hdl.handle.net/2117/89946
Access Level:acceso abierto
Palabra clave:Mathematical statistics
Regression analysis
Telecom networks
Data analysis
Traffic modelling
Statistics
Machine learning
Estadística matemàtica
Classificació AMS::62 Statistics::62J Linear inference, regression
Àrees temàtiques de la UPC::Matemàtiques i estadística::Estadística matemàtica
Descripción
Sumario:The objective of this project is to provide traffic models based on new services characteristics. Specifically, we focus on modelling the traffic between origin-destination node pairs (also known as OD pairs) in a telecom network. Two use cases are distinguished: i) traffic generation in the context of simulation, and ii) traffic modelling for prediction in the context of big-data backed telecom cloud systems. To this aim, several machine learning and statistical models and technics are studied and combined in order to find the best approach for every use case. To evaluate the applicability of selected models, we integrated them in an OMNeT++ network simulator whose implementation follows the Big Data analytics architecture for the telecom cloud.