Traffic modeling and optimization in datacenters with graph neural network

Abstract Traffic Optimization (TO) is a well-known and established topic in datacenters with the fundamental goal of operating networks efficiently. Traditional TO heuristics may suffer from performance penalty as it mismatches actual traffic, while Artificial Intelligence (AI) which has undergone a renaissance recently is gradually being applied to the network optimization and has shown excellent advantages. However, the current AI technologies (e.g., DGN, DRL, DBA, etc.) have difficulty in adapting to the dynamic and variable characteristics of the network due to their lack of generalization ability, which limits the development of intelligent networks. As Graph Neural Network (GNN) can support relational reasoning and combinatorial generalization, we research how to model and optimize traffic in datacenters with GNN in this paper. First, we proposed a GNN model for reasoning Flow Completion Time (FCT), which is able to provide accurate estimation of never-seen network states. Then we designed a GNN-based optimizer for TO, which can be used to in flow routing, flow scheduling and topology management. Finally, the experimental results verify that the GNN model has a high inference accuracy, and the GNN-based optimizer can significantly reduce the average / p10 (the 10th percentile) FCT. Therefore, GNN has a great potential in network modeling and optimization, and has a wide range of applications.

[1]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[2]  Razvan Pascanu,et al.  Interaction Networks for Learning about Objects, Relations and Physics , 2016, NIPS.

[3]  Khaled Elleithy,et al.  Efficient Actor Recovery Paradigm for Wireless Sensor and Actor Networks , 2017, Sensors.

[4]  Tim Moors,et al.  A Flow-Level Extension to OMNeT++ for Long Simulations of Large Networks , 2017, IEEE Communications Letters.

[5]  Feng Liu,et al.  AuTO: scaling deep reinforcement learning for datacenter-scale automatic traffic optimization , 2018, SIGCOMM.

[6]  Keqiang He,et al.  Presto: Edge-based Load Balancing for Fast Datacenter Networks , 2015, SIGCOMM.

[7]  J. Klafter,et al.  Paretian Poisson Processes , 2008 .

[8]  Ankit Singla,et al.  OSA: An Optical Switching Architecture for Data Center Networks With Unprecedented Flexibility , 2012, IEEE/ACM Transactions on Networking.

[9]  Albert G. Greenberg,et al.  VL2: a scalable and flexible data center network , 2009, SIGCOMM '09.

[10]  Alexander M. Rush,et al.  Character-Aware Neural Language Models , 2015, AAAI.

[11]  Hongyi Wu,et al.  A New Data Transmission Strategy in Mobile D2D Networks—Deterministic, Greedy, or Planned Opportunistic Routing? , 2017, IEEE Transactions on Vehicular Technology.

[12]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[13]  Shahenda Sarhan,et al.  A novel hybrid of Shortest job first and round Robin with dynamic variable quantum time task scheduling technique , 2017, Journal of Cloud Computing.

[14]  Jiao Zhang,et al.  Flow distribution-aware load balancing for the datacenter , 2017, Comput. Commun..

[15]  Francesco Musumeci,et al.  Elastic operations in federated datacenters for performance and cost optimization , 2014, Comput. Commun..

[16]  Joseph E. Ford,et al.  A Scalable, Partially Configurable Optical Switch for Data Center Networks , 2017, Journal of Lightwave Technology.

[17]  Xiaoming Fu,et al.  A Survey on Virtual Machine Migration: Challenges, Techniques, and Open Issues , 2018, IEEE Communications Surveys & Tutorials.

[18]  Wei Bai,et al.  Information-Agnostic Flow Scheduling for Commodity Data Centers , 2015, NSDI.

[19]  Nei Kato,et al.  Routing or Computing? The Paradigm Shift Towards Intelligent Computer Network Packet Transmission Based on Deep Learning , 2017, IEEE Transactions on Computers.

[20]  Florin Ciucu,et al.  Perspectives on network calculus: no free lunch, but still good value , 2012, CCRV.