ABSTRACT
We are currently witnessing a rapidly increasing number of
mobile users utilising the Transmission Control Protocol (TCP) over wireless
networks for accessing Internet services such as video streaming services over
HTTP. TCP has been designed for wireline networks and its shortcomings; such as
throughput degradation due to random losses and intermittent connectivity, have
been the subject of a large volume of research investigations over the last few
years. In this thesis, a set of techniques are proposed to enhance the
performance of the end-to-end wireless communications using TCP as a transport
layer protocol. The proposed set of technique is a Cross-Layer solution that
integrates some information from the link layer as the value of signal strength
and noise with the TCP packet loss control mechanism of TCP connections. In the
design of such a smart transport layer such as on a multi-hop connexion,
important issues are raised such as deciding which value of signal strength and
noise to choose, which threshold we need to set and how to estimate the best
RTT value. Through an extensive series of simulations on the performance of the
proposed techniques while focusing on the variables that affect the experience
of the end-user, the end-to-end throughput that a TCP flow can accomplish was
considered.
As a next step in working on TCP performances over wireless
network, we have chosen to apply our first contribution with the HTTP adaptive
streaming (HAS) to increase the users experience QoE. The HTTP adaptive
streaming (HAS) is a streaming video technique widely used over the Internet
for Video on Demand (VoD) and Live streaming services. It employs Transmission
Control Protocol (TCP) as transport protocol and it splits the original video
inside the server into segments of same duration, called "chunks", that are
transcoded into multiple quality levels. In this thesis, we proposed to
integrate the user feedback and his terminal parameters (i.e. resolution,
screen, battery) on the adaptation process by using the TCP parameters tuning.
To estimate the user satisfaction we used the mean opinion score (MOS) of the
users which is a score out of five points that the user gives to express his
satisfaction towards the proposed set of techniques. Compared to other adaptive
video streaming solutions, the emulation results show the extent to which our
proposed solution can increase the user experience (satisfaction) on this kind
of service.
Keywords: TCP, Congestion, Signal strength,
Noise, MANET, RTT, QoE, QoS, Video streaming, Terminal device, Multimedia,
Reinforcement learning.
vi
TABLE DES MATIERES
DEDICACES ii
REMERCIEMENTS ...iii
RESUME iv
ABSTRACT . .v
TABLE DES MATIERES vi
LISTE DES FIGURES vii
LISTE DES TABLES .viii
LISTE DES ACRONYMES ...ix
PUBLICATIONS ET CONFERENCES ...X
INTRODUCTION GENERALE 1
1. CONTEXTE ET PROBLEMATIQUE DE LA THESE
1
2. CONTRIBUTION ET STRUCTURE DE LA THESE
..3
CHAPITRE I : CONCEPTS DE BASE
|
1. GENERALITES SUR LES RESEAUX SANS FIL
|
...6
|
|
1.1. Introduction
|
6
|
|
1.2. Définition
|
.6
|
|
1.3. Les différentes technologies des réseaux
sans fil
|
7
|
|
1.4. Différents normes de réseaux sans fil
|
8
|
|
1.5. Spécificités des réseaux sans fil
|
.10
|
|
1.5.1. Spécificité physique
|
..10
|
|
1.5.2. Erreur du canal
|
10
|
|
1.5.3. Contention du Médium et collision
|
..11
|
vi
1.5.4. Mobilité 12
1.5.5. Spécificité du routage 13
1.5.6. Congestion 13
1.5.7. Considérations énergétiques 14
1.6. Les réseaux sans fils ad hoc mobiles 14
1.7. Caractéristiques des réseaux ad hoc 15
2. PROTOCOLE DE TRANSPORT 17
2.1. Introduction ..17
2.2. Le protocole TCP 18
2.2.1. Caractéristiques et fonctionnement
général .18
3.FONCTIONS DE CONTROLE DE CONGESTION
3.1. La phase slow-start (démarrage lent) 22
3.2. Congestion avoidance (évitement de congestion)
23
3.3. L'algorithme Additive Increase and Multiplicative
Decrease (AIMD) 23
3.4. La reprise sur erreur . 23
3.4.1. Fast retransmit (retransmission rapide) 24
3.4.2. Fast-recovey (recouvrement rapide) . 24
3.4.3. Selective Acknowledgment (SACK) 24
4.VARIANTES DE TCP 25
4.1. TCP Tahoe 25
4.2. TCP Reno 25
4.3. TCP New Reno 26
4.4. TCP Vegas 26
4.5. TCP Westwood+ 26
4.6. TCP SACK 27
5.TCP ET LES RESEAUX SANS FIL 27
5.1. Problèmes de TCP dans les réseaux ad hoc
mobiles ...28
6.LE PROTOCOLE HTTP 30
7.LES SERVICES DE VIDEO STREAMING ..
32
7.1. Introduction 32
vi
7.2. Le streaming vidéo 32
7.3. Le système de streaming multimédia 33
8.LES DIFFERENTES TECHNOLOGIES DE STREAMING
34
8.1. Le streaming en direct et le streaming stocké
34
8.1.1. Streaming adaptatif 35
8.1.2. Le streaming non adaptatif 35
9. Conclusion ..36
| 
