Espace membre

Cet espace est dédié aux étudiants, aux enseignants et au personnel administratif de l'école

Valider

Mot de passe oublié?

Actualités de l'école

19/12/2019

Soutenance de doctorat de Amal Kammoun

Soutenance de thèse de doctorat de Amal Kammoun le 19/12/2019 à 10H00 , à l'amphi Ibn Khaldoun de SUP'COM


Intitulé :    SDN/NFV-based Network Slicing Management


Présentée par : Amal Kammoun

JURY


President

Mme Lamia Chaari        

Professor, ISIMS, SFAX University, Tunisia

 

 

 

Reviewers

M. Michelle Sibilla            

Professor, Université Toulouse III, France

  Mme Kaouther Sethom
MC HDR, ENICarthage, Tunisia
     

Examiner

M. Tijani Chahed       Professor, Télécom SudParis, Paris, France
     

Supervisor

M. Nabil Tabbane          

Professor, SUP'COM, Tunisia

 

 

 

Supervisor

Mme Gladys Diaz

MC HDR, Université Paris 13, France

Abstract

Fifth Generation of Mobile Networks (5G) is proposed to address the future network context. This context is characterized by an everywhere-connected society where not only persons will be connected to the network but also clothes, vehicles and smart objects. Therefore, 5G has to serve an expanding number of users and terminals and thus an exponentially increasing traffic. This traffic will be generated from several emerging services such as e-health, autonomous driving, IoT, etc. These services request different and stringent constraints mainly in terms of latency, reliability, and availability.

For instance, for the smart city applications, current network architectures are not able to handle neither the increased number of sensors and IoT devices nor the large amount of data exchanged over the network. 5G networks try to cope with the limitations of current network implementations by proposing a new system aiming to meet new challenges. Indeed, unlike previous technologies, 5G will not only enhance the network system but will also provide an end-to-end infrastructure that will support emergent services and respond to stringent user requirements. The key concepts for the 5G vision are the Software Defined Networking (SDN), Network Function Virtualization (NFV) and Network Slicing technologies. Those paradigms allow the network to provide services for various scenarios under different requirements. They permit to achieve higher performance and flexibility for the network through the introduction of the network programming and the one size per service approach instead of the traditional one size fits all approach.

In this context, this thesis aims first to propose a new architecture for network slices provisioning and management. We propose the multi-level Delegation Architecture for Network Slicing Orchestration (DANSO) architecture which considers the SDN, NFV and network slicing technologies in order to present a programmable and flexible framework for services provisioning. The second aim is the optimization of the process of the network slices creation and the admission control of users’ request. For this purpose, we propose heuristic algorithms in order to either map the users’ demands to existing slices or to create a new network slices. Our algorithms consider the reliability, availability and latency requirements as well as the offered quality by the underlying infrastructure. The third aim is related to the management aspect. In fact, we interest on the management of sudden events that occur in the slice during its running time. In this regard, we study the congestion of the slices and users’ mobility events. We propose a fuzzy logic-based algorithm that considers the actual and the predicted load state of the slice in order to perform auto-scaling actions. The future load values are determined using the Support Vector Regression (SVR) technique. Finally we interest on the problem of vertical handover management in the context of network slices. We propose an algorithm that decides when to perform the handover and selects the target slice.

 

Keywords

SDN, NFV, Optimization, Management, Network Slicing, Reliability, Availability, Prediction, Handover, Mobility, Fuzzy Logic