Improving Traffic Splitting Accuracy based on Flow Size Prediction

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K17730
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 60060:Information network-related
• Research Institution: The University of Electro-Communications
• Principal Investigator: Yamaki Hayato 電気通信大学, 大学院情報理工学研究科, 准教授 (20782197)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: トラフィック分割

Outline of Final Research Achievements

In this study, we evaluated methods for splitting communication traffic with higher accuracy than before and the impact of such methods on the performance of communication applications. While traffic partitioning at the packet level achieves high accuracy, it disrupts the arrival order of packets at the receiver side and requires re-alignment processing. Therefore, we proposed a method to divide traffic at the packet level without disturbing the arrival order of packets as much as possible. We also focus on link aggregation (LAG) and multipath routing as communication applications, and show that traffic partitioning with high accuracy can improve throughput compared to conventional methods.

Free Research Field

ネットワークアーキテクチャ

Academic Significance and Societal Importance of the Research Achievements

近年，動画などの高容量コンテンツの配信やIoT（Internet of Things）における超多ノードでの通信など，ネットワークの大容量化・高スループット化は必要不可欠となっている．本研究が着目するトラフィック分割は，このような要求を解決する一方法であるリンク集約やマルチパスルーティングの性能に直結する要素であり，実際に本研究ではこれらのアプリケーションにおいてスループットが向上できることを示している．本研究は今後の通信分野の発展に貢献する高い意義を有すると言える．