Skip to Main content Skip to Navigation
Conference papers

When Deep Learning meets Web Measurements to infer Network Performance

Imane Taibi 1, 2 Yassine Hadjadj-Aoul 1 Chadi Barakat 2
1 DIONYSOS - Dependability Interoperability and perfOrmance aNalYsiS Of networkS
IRISA-D2 - RÉSEAUX, TÉLÉCOMMUNICATION ET SERVICES, Inria Rennes – Bretagne Atlantique
2 DIANA - Design, Implementation and Analysis of Networking Architectures
CRISAM - Inria Sophia Antipolis - Méditerranée
Abstract : Web browsing remains one of the dominant applications of the internet, so inferring network performance becomes crucial for both users and providers (access and content) so as to be able to identify the root cause of any service degradation. Recent works have proposed several network troubleshooting tools, e.g, NDT, MobiPerf, SpeedTest, Fathom. Yet, these tools are either computationally expensive, less generic or greedy in terms of data consumption. The main purpose of this work is to leverage passive measurements freely available in the browser and machine learning techniques (ML) to infer network performance (e.g., delay, bandwidth and loss rate) without the addition of new measurement overhead. To enable this inference, we propose a framework based on extensive controlled experiments where network configurations are artificially variedand the Web is browsed, then ML is applied to build models that estimate the underlying network performance. In particular, we contrast classical ML techniques (such as random forest) to deep learning models trained using fully connected neural networks and convolutional neural networks (CNN). Results of our experiments show that neural networks have a higher accuracy compared to classical ML approaches. Furthermore, the model accuracy improves considerably using CNN.
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.inria.fr/hal-02358004
Contributor : Chadi Barakat <>
Submitted on : Monday, November 11, 2019 - 10:19:58 AM
Last modification on : Tuesday, February 25, 2020 - 8:08:10 AM
Document(s) archivé(s) le : Wednesday, February 12, 2020 - 2:20:18 PM

File

1570571952.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02358004, version 1

Citation

Imane Taibi, Yassine Hadjadj-Aoul, Chadi Barakat. When Deep Learning meets Web Measurements to infer Network Performance. CCNC 2020 - IEEE Consumer Communications & Networking Conference, Jan 2020, Las Vegas, United States. pp.1-6. ⟨hal-02358004⟩

Share

Metrics

Record views

106

Files downloads

565