Multipath TCP (MPTCP) is considered as a solution capable of addressing the growing demand for bandwidth. However, the existing MPTCP mechanisms make flow scheduling based on coarse-grained end-to-end network states, which prevents MPTCP from better aggregating the bandwidth of multiple paths. Besides, link overlapping may occur between different MPTCP connections, which results in multiple subflows competing for bandwidth of the shared link. In this paper, we propose GFlow, a Graph Neural Network (GNN) based Deep Reinforcement Learning (DRL) algorithm, to make optimal flow scheduling for multipath transmission with link overlapping. Specifically, we formulate the flow scheduling problem as a problem of maximizing overall throughput by taking both bottleneck bandwidth and shared bandwidth into consideration. To support accurate network state perception, GFlow utilizes In-band Network Telemetry (INT) to collect real-time and fine-grained network states. Taking these states as input, the DRL agent with GNN integrated fully learns the relationships among links, paths (subflows), and MPTCP connections. In this way, GFlow is able to make optimal flow scheduling decisions according to the network states. We build a P4-based multipath transmission system and carry out extensive experiments to evaluate the performance of GFlow. The results show that GFlow outperforms the baseline multipath transmission mechanism in both homogeneous scenario and heterogeneous scenario, improving the average overallthroughput while reducing the average round trip time (RTT).

Du Chen

Weiting Zhang

Deyun Gao

Dong Yang

Hongke Zhang

IEEE Transactions on Network Science and Engineering

拟议的期刊，称为 IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING (TNSE)，致力于及时发表同行评议的技术文章，这些文章涉及网络科学的理论和应用，以及构成系统的元素之间的互连。网络。特别是，IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 发表了关于在基础层面理解、预测和控制网络结构和行为的文章。涵盖的网络类型包括物理或工程网络、信息网络、生物网络、语义网络、经济网络、社交网络和生态网络。该杂志旨在发现管理网络结构、网络功能和网络行为的共同原则，寻求关于理解、预测和控制网络结构和行为的文章。 IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 的另一个跨学科重点是不同类型网络之间的相互作用和共同演化。

The IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING (TNSE) is a journal dedicated to the timely publication of peer-reviewed technical articles that address the theory and applications of network science, as well as the interconnections among the elements that constitute network systems. This journal particularly focuses on research aimed at understanding, predicting, and controlling network structures and behaviors at a fundamental level. The types of networks covered include physical or engineering networks, information networks, biological networks, semantic networks, economic networks, social networks, and ecological networks. The journal aims to discover common principles for managing network structures, functions, and behaviors, seeking articles that contribute to the understanding, prediction, and control of network structures and behaviors. Another interdisciplinary focus of the IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING is the interactions and co-evolution between different types of networks.

IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING (TNSE) 是一本致力于及时发表同行评审的技术文章的期刊，涵盖网络科学的理论与应用，以及构成网络系统的元素之间的相互连接。该期刊特别关注在基础层面上理解、预测和控制网络结构与行为的研究。涵盖的网络类型包括物理或工程网络、信息网络、生物网络、语义网络、经济网络、社交网络和生态网络。期刊旨在发现管理网络结构、功能和行为的共同原则，并寻求关于理解、预测和控制网络结构与行为的研究文章。IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING 还特别关注不同类型网络之间的相互作用与共同演化。

GFlow: GNN-Based Optimal Flow Scheduling for Multipath Transmission with Link Overlapping

GFlow, a Graph Neural Network based Deep Reinforcement Learning (DRL) based Deep Reinforcement Learning (DRL) algorithm, is proposed to make optimal flow scheduling for multipath transmission with link overlapping, and outperforms the baseline multipath transmission mechanism in both homogeneous scenario and heterogeneous scenario.

Category	Quartile
计算机科学	2
计算机科学, 数学跨学科应用	1
计算机科学, 工程综合	2

GFlow: GNN-Based Optimal Flow Scheduling for Multipath Transmission with Link Overlapping

ivySCI AI Smartly Parses PDF, Answers Researchers' Questions, and Helps You Understand Papers in Seconds

Journal Info

Category	Quartile
MATHEMATICS, INTERDISCIPLINARY APPLICATIONS	1