Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
DEAR: DRL Empowered Actor-Critic ScheduleR for Multipath QUIC Under 5G/B5G Hybrid Networks
BITS Pilani, India.
BITS Pilani, India.
Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science (from 2013). University of Oslo, Norway.ORCID iD: 0000-0001-5800-2779
2024 (English)In: Advanced Information Networking and Applications: Proceedings of the 38th International Conference on Advanced Information Networking and Applications (AINA-2024), Volume 1 / [ed] Leonard Barolli, Springer, 2024, Vol. 199, p. 103-113Conference paper, Published paper (Refereed)
Abstract [en]

Recently, the Internet has experienced a substantial increase in the use of bandwidth-intensive applications due to the introduction of fifth generation (5G) and beyond 5G (B5G) systems. Empirical evidence has shown that multipath transport layer protocols such as multipath TCP (MPTCP) and multipath QUIC (MPQUIC) are successful in addressing the increasing need for higher bandwidth in the existing Internet infrastructure. Nevertheless, multipath schedulers still face difficulties in efficiently handling significant amounts of variability in diverse network scenarios. This paper introduces, Deep reinforcement learning (DRL) Empowered Actor-critic scheduleR, DEAR, a method designed for multipath QUIC in 5G/B5G hybrid networks. DEAR is developed utilising a DRL based actor critic methodology. This approach significantly improves the decision-making abilities of the scheduler in various rapidly changing network scenarios. We conducted experiments with the DEAR scheduler in several network settings, encompassing networks with rapidly fluctuating bandwidth, networks with rapid and short-term fluctuations, and networks experiencing progressive outages. We also conducted tests on the DEAR algorithm using the Lumos5G dataset, which consists of real network traces from two distinct service providers. We have performed a comparative analysis of DEAR with other state-of-the-art multipath schedulers, including another scheduler based on RL, Peekaboo, and other non-RL rule-based schedulers such as round robin (RR), earliest completion first (ECF), blocking estimation (BLEST), and minimum round trip time (minRTT). Our evaluation demonstrates DEAR’s superior performance compared to existing algorithms. In scenarios with fast-changing bandwidth, DEAR outperforms rule-based schedulers by 42.30% and Peekaboo by 17.77%. Similarly, in networks with fast short-scale variations, DEAR achieves gains of 22.22% over rule-based schedulers and 6.06% over Peekaboo. Moreover, in networks facing progressive outage and recovery, DEAR showcases gains of 13.90% over rule-based schedulers and 8.79% over Peekaboo. 

Place, publisher, year, edition, pages
Springer, 2024. Vol. 199, p. 103-113
Series
Lecture Notes on Data Engineering and Communications Technologies, ISSN 2367-4512, E-ISSN 2367-4520 ; LNDECT,volume 199
Keywords [en]
5G/B5G Networks, MPQUIC, Multipath Networking, Schedulers
National Category
Computer Sciences
Research subject
Computer Science
Identifiers
URN: urn:nbn:se:kau:diva-99741DOI: 10.1007/978-3-031-57840-3_10Scopus ID: 2-s2.0-85191338992ISBN: 978-3-031-57839-7 (print)ISBN: 978-3-031-57840-3 (electronic)OAI: oai:DiVA.org:kau-99741DiVA, id: diva2:1864144
Conference
38th International Conference on Advanced Information Networking and Applications,Kitakyushu, Japan, April 17-19, 2024.
Available from: 2024-06-03 Created: 2024-06-03 Last updated: 2024-06-03Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Alay, Özgü

Search in DiVA

By author/editor
Alay, Özgü
By organisation
Department of Mathematics and Computer Science (from 2013)
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 45 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf