<i>MTS‐PolKA</i>: Weighted Multipath Traffic Splitting With Source Routing for Elephant and Mice Flows

ABSTRACT Modern datacenter networks rely on multi‐root tree topologies, to ensure efficient and reliable connectivity. However, traditional traffic engineering solutions like equal cost multiple path (ECMP) and weighted‐cost multipath (WCMP) often fall short in addressing dynamic and heterogeneous traffic conditions. Specifically, these approaches struggle with asymmetric topologies and the allocation of elephant flows, which require more granular and agile traffic management. This paper introduces MTS‐PolKA ( Multipath Traffic Split Polynomial Key‐based Architecture ), a novel traffic splitting mechanism designed to enhance traffic engineering in datacenter networks. MTS‐PolKA enables packet‐level traffic distribution across multiple paths by embedding a source routing label in the packet header that dictates the per‐flow traffic division profile. Unlike existing methods that require reconfiguring network tables at each switch, MTS‐PolKA enables dynamic adjustments across all switches in the path through simple packet header modifications, significantly enhancing the agility and efficiency of traffic management. The implementation of MTS‐PolKA on programmable switches using the P4 language, along with the residue number system (RNS) and M‐PolKA architecture, showcases its potential through experiments conducted in the Mininet emulator. The results highlight the ability of MTS‐PolKA to maintain flow stability, explore network multipaths, and enable rapid reconfigurations of traffic division profiles, enhancing the performance and efficiency of modern datacenter networks.

Authors

Institutions

Publication Details

Journal
International Journal of Network Management
Published
2025-04-16
DOI
https://doi.org/10.1002/nem.70016
Primary Topic
Advanced Memory and Neural Computing
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

MTS‐PolKA: Weighted Multipath Traffic Splitting With Source Routing for Elephant and Mice Flows

Gilmar L. Vassoler, Rodolfo da Silva Villaça, Cristina K. Dominicini, Rafael S. Guimarães et al.
International Journal of Network Management
Advanced Memory and Neural Computing
article

MTS‐PolKA: Weighted Multipath Traffic Splitting With Source Routing for Elephant and Mice Flows

Gilmar L. Vassoler, Rodolfo da Silva Villaça, Cristina K. Dominicini, Rafael S. Guimarães, Domingos Jose P. Paraiso, Isis Oliveira, Giancarlo O. dos Santos
article en

Abstract

ABSTRACT Modern datacenter networks rely on multi‐root tree topologies, to ensure efficient and reliable connectivity. However, traditional traffic engineering solutions like equal cost multiple path (ECMP) and weighted‐cost multipath (WCMP) often fall short in addressing dynamic and heterogeneous traffic conditions. Specifically, these approaches struggle with asymmetric topologies and the allocation of elephant flows, which require more granular and agile traffic management. This paper introduces MTS‐PolKA ( Multipath Traffic Split Polynomial Key‐based Architecture ), a novel traffic splitting mechanism designed to enhance traffic engineering in datacenter networks. MTS‐PolKA enables packet‐level traffic distribution across multiple paths by embedding a source routing label in the packet header that dictates the per‐flow traffic division profile. Unlike existing methods that require reconfiguring network tables at each switch, MTS‐PolKA enables dynamic adjustments across all switches in the path through simple packet header modifications, significantly enhancing the agility and efficiency of traffic management. The implementation of MTS‐PolKA on programmable switches using the P4 language, along with the residue number system (RNS) and M‐PolKA architecture, showcases its potential through experiments conducted in the Mininet emulator. The results highlight the ability of MTS‐PolKA to maintain flow stability, explore network multipaths, and enable rapid reconfigurations of traffic division profiles, enhancing the performance and efficiency of modern datacenter networks.

International Journal of Network ManagementVol. 35(3)
Universidade Federal do Espírito Santo (BR)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Ministério da Ciência, Tecnologia e Inovação, Fundação de Amparo à Pesquisa e Inovação do Espírito Santo
Openalex Percentile: Top 94%
Advanced Memory and Neural Computing
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.