DC Field | Value | Language |
dc.contributor.author | BOUFELDJA, MAnel | - |
dc.contributor.author | SAFARI, MAouadda | - |
dc.date.accessioned | 2023-10-15T13:28:21Z | - |
dc.date.available | 2023-10-15T13:28:21Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | https://repository.esi-sba.dz/jspui/handle/123456789/511 | - |
dc.description | Encadreur : Dr. Zouaoui Chakib / Co-Encadreur : Dr. Manel Baba-Ahmed | en_US |
dc.description.abstract | Abstract :
An essential element in MPTCP is the congestion control algorithm (CCA), which plays a
vital role in regulating the transmission of data across multiple subflows (SFs). However,
there are two primary challenges associated with MPTCP CCAs. Firstly, MPTCP flows
need to possess an edge over single-path flows. Secondly, MPTCP flows should exhibit
fairness, meaning that SFs sharing a common bottleneck should utilize a bandwidth comparable
to that of a single-path flow.
Despite the creation of multiple MPTCP congestion control algorithms (CCAs),
none of them have managed to overcome these challenges in all scenarios. In response to
this, Google introduced the Bottleneck Bandwidth and Round-trip Time (BBR), a novel
CCA designed for single-path TCP. BBR utilizes a network model to achieve exceptional
throughput while minimizing delays.
To leverage the high performance of BBR, in this work, we adapted this algorithm
to be coupled and operate within MPTCP. To achieve this objective, we utilized
the principle of the Linked Increase Algorithm (LIA).
This work has successfully published on GitHub SAFARI and BOUFELDJA [2023] as a
public repository. The repository serves as a comprehensive demonstration of the project’s
code, and resources, making it accessible to a wider audience of developers and researchers.***
Résumé
Un ´el´ement essentiel du MPTCP est l’algorithme de contrˆole de la congestion (CCA),
qui joue un rˆole vital dans la r´egulation de la transmission des donn´ees `a travers de
multiples sous-flux (SF). Cependant, les CCA du MPTCP posent deux probl`emes majeurs.
Premi`erement, les flux MPTCP doivent poss´eder un avantage sur les flux `a chemin
unique. Deuxi`emement, les flux MPTCP doivent ˆetre ´equitables, ce qui signifie que les
FS partageant un goulot d’´etranglement commun doivent utiliser une largeur de bande
comparable `a celle d’un flux `a chemin unique.
Malgr´e la cr´eation de plusieurs algorithmes de contrˆole de la congestion (CCA)
pour MPTCP, aucun d’entre eux n’a r´eussi `a surmonter ces difficult´es dans tous les
sc´enarios, aucun d’entre eux n’a r´eussi `a surmonter ces d´efis dans tous les sc´enarios.
C’est pourquoi Google a introduit l’algorithme BBR (Bottleneck Bandwidth and Roundtrip
Time), Google a introduit le Bottleneck Bandwidth and Round-trip Time (BBR), un
nouvel ACC con¸cu pour le TCP `a chemin unique. BBR utilise un mod`ele de r´eseau pour
atteindre un d´ebit exceptionnel tout en minimisant les d´elais.
Pour b´en´eficier des hautes performances de BBR, dans ce travail, nous avons
adapt´e cet algorithme pour le coupler et le faire fonctionner dans MPTCP. Afin d’atteindre
cet objectif, nous avons utilis´e le principe de l’algorithme Linked Increase (LIA).
Ce travail a ´et´e publi´e avec succ`es sur GitHub SAFARI and BOUFELDJA [2023] en
tant que r´ef´erentiel public. Le r´ef´erentiel sert de d´emonstration compl`ete du code et
des ressources du projet, le rendant accessible `a un large public de d´eveloppeurs et de
chercheurs. | en_US |
dc.language.iso | en | en_US |
dc.subject | BBR | en_US |
dc.subject | MPTCP | en_US |
dc.subject | TCP | en_US |
dc.subject | LIA | en_US |
dc.subject | Congestion Control | en_US |
dc.title | Design and implementation of a new optimized coupled BBR | en_US |
dc.type | Thesis | en_US |
Appears in Collections: | Ingénieur
|