File(s) not publicly available
Fair and Efficient Distributed Edge Learning With Hybrid Multipath TCP
journal contribution
posted on 2023-02-20, 04:20 authored by Shiva PokhrelShiva Pokhrel, Jinho ChoiJinho Choi, A WalidThe bottleneck of distributed edge learning (DEL) over wireless has shifted from computing to communication, primarily the aggregation-averaging (Agg-Avg) process of DEL. The existing transmission control protocol (TCP)-based data networking schemes for DEL are application-agnostic and fail to deliver adjustments according to application layer requirements. As a result, they introduce massive excess time and undesired issues such as unfairness and stragglers. Other prior mitigation solutions have significant limitations as they balance data flow rate from workers across paths but often incur imbalanced backlogs when the paths exhibit variance, causing stragglers. To facilitate a more productive DEL, we develop a hybrid multipath TCP (MPTCP) by combining model-based and deep reinforcement learning (DRL) based MPTCP for DEL that strives to realize quicker iteration of DEL and better fairness (by ameliorating stragglers). Hybrid MPTCP essentially integrates two radical TCP developments: i) successful existing model-based MPTCP control strategies and ii) advanced emerging DRL-based techniques, and introduce a novel hybrid MPTCP data transport for easing the communication of Agg-Avg process. Extensive emulation results demonstrate that the proposed hybrid MPTCP can overcome excess time consumption and ameliorate the application layer unfairness of DEL effectively without injecting additional inconstancy and stragglers.
History
Journal
IEEE/ACM Transactions on NetworkingVolume
PPPublisher DOI
ISSN
1063-6692eISSN
1558-2566Language
EnglishPublication classification
C1 Refereed article in a scholarly journalIssue
99Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INCUsage metrics
Categories
Keywords
BANDWIDTH ALLOCATIONComputer ScienceComputer Science, Hardware & ArchitectureComputer Science, Theory & MethodsCONGESTION CONTROLData parallelismdeep reinforcement learningDESIGNdistributed edge learningEngineeringEngineering, Electrical & Electronicmultipath TCPNETWORKSScience & TechnologyTechnologyTelecommunicationsCommunications Technologies not elsewhere classifiedElectrical and Electronic Engineering not elsewhere classifiedDistributed Computing
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC