logo EDITE Jordan AUGE
Jordan AUGE
État académique
Thèse soutenue le 2014-11-27
Sujet: Garanties de performance pour les flots IP dans l'architecture Flow-Aware Networking
Direction de thèse:
Encadrement de thèse:
Ellipse bleue: doctorant, ellipse jaune: docteur, rectangle vert: permanent, rectangle jaune: HDR. Trait vert: encadrant de thèse, trait bleu: directeur de thèse, pointillé: jury d'évaluation à mi-parcours ou jury de thèse.
Productions scientifiques
TopHat: supporting experiments through measurement infrastructure federation
Researchers use the PlanetLab testbed for its ability to host experimental applications in realistic conditions over the public best-effort internet. Such applications form overlays whose performance is affected by the underlying topology and its evolution. While several topology information services have been proposed for PlanetLab, the TopHat system that we describe here fills a special niche. It is designed to support the entire lifecycle of an experiment: from setup, through run time, to retrospective analysis. TopHat does so in a new way, by drawing upon excellent, proven third party services, notably the Dimes and Etomic measurement infrastructures, for specialized measurements. TopHat has been developed as the active measurement component of PlanetLab Europe, the flagship testbed of the OneLab experimental facility. It is part of OneLab's larger effort to pioneer the federation of previously independent testbeds and measurement systems in order to provide a diverse global scale environment for Future Internet research.
Proc. TridentCom, Berlin, Germany 2010
Violation of interdomain routing assumptions
We challenge a set of assumptions that are frequently used to model interdomain routing in the Internet by confronting them with routing decisions that are actually taken by ASes, as revealed through publicly available BGP feeds. Our results quantify for the first time the extent to which such assumptions are too simple to model real-world Internet routing policies. This should introduce a note of caution into future work that makes these assumptions and should prompt attempts to find more accurate models.
PAM'2014 Passive and Active Measurement (PAM)conference proceeding 2014-03-10
Tools to foster a global federation of testbeds
A global federation of experimental facilities in computer networking is being built on the basis of a thin waist, the Slice-based Federation Architecture (SFA), for managing testbed resources in a secure and efficient way. Its success will depend on the existence of tools that allow testbeds to expose their local resources and users to browse and select the resources most appropriate for their experiments. This paper presents two such tools. First, SFAWrap, which makes it relatively easy for a testbed owner to provide an SFA interface for their testbed. Second, MySlice, a tool that allows experimenters to browse and reserve testbed resources via SFA, and that is extensible through a system of plug-ins. Together, these tools should lower the barriers to entry for testbed owners who wish to join the global federation.
Computer Networks, Special Issue on Future Internet Testbedsarticle in peer-reviewed journal 2014-01-03
Measurement-based admission control for flow-aware implicit service differentiation
It has previously been shown that the combined use of fair queuing and admission control would allow the Internet to provide satisfactory quality of service for both streaming and elastic flows without explicitly identifying traffic classes. In this paper we discuss the design of the required measurement based admission control (MBAC) scheme. The context is different to that of previous work on MBAC in that there is no prior knowledge of flow characteristics and there is a twofold objective: to maintain adequate throughput for elastic flows and to ensure low packet latency for any flow whose peak rate is less than a given threshold. In the paper we consider the second objective assuming realistically that most elastic and streaming flows are rate limited. We propose a MBAC algorithm and evaluate its performance by simulation under different stationary traffic mixes and in a flash crowd scenario. The algorithm is shown to offer a satisfactory compromise between flow performance and link utilization.
Proceeding of 23rd International Teletraffic Congress (ITC 2011) ITC 23 : 23rd International Teletraffic Congressconference proceeding 2011-09-09
A Statistical Bandwidth Sharing Perspective on Buffer Sizing
International audience
The issue of buffer sizing is rightly receiving increasing at-tention with the realization that the bandwidth delay product rule-of-thumb is becoming unsustainable as link capacity continues to grow. In the present paper we examine this issue from the light of our understand-ing of traffic characteristics and the performance of statistical bandwidth sharing. We demonstrate through simple analytical models coupled with the results of ns2 simulations that, while a buffer equivalent to the band-width delay product is certainly unnecessary, the recently advocated re-duction to a few dozen packets is too drastic. The required buffer size depends significantly on the peak exogenous rate of multiplexed flows.
20th International Teletraffic Congress https://hal.archives-ouvertes.fr/hal-01092848 20th International Teletraffic Congress, Jun 2007, Ottawa, Canada. <10.1007/978-3-540-72990-7_38>Conference papers 2007-06-17
Buffer sizing for elastic traffic
International audience
Two areas of active research are likely to have a major impact on the future performance of the Internet. These are firstly, the evaluation of the buffer size required to ensure fair, stable and efficient sharing of link bandwidth and secondly, the proposition of new congestion control algorithms, more suitable for increasingly high link speeds than the current version of TCP. In this paper, we re-examine propositions for a drastic reduction in buffer size in the light of our understanding of the nature of Internet traffic at flow level. We demonstrate through simulation that the trade-off between throughput and delay is not as favourable as previously predicted. We also illustrate the significant gains in this context that would accrue from the implementation of per-flow fair scheduling in router queues.
NGI2006, 2nd Conference on Next Generation Internet Design and Engineering https://hal.archives-ouvertes.fr/hal-01092865 NGI2006, 2nd Conference on Next Generation Internet Design and Engineering, Apr 2006, Valencia, Spain. pp.33 - 40, <10.1109/NGI.2006.1678220>Conference papers 2006-04-03
Thèse: "Garanties de performances pour les flots IP dans l'architecture Flow-Aware Networking"
Soutenance: 2014-11-27