Is Machine Learning Ready for Traffic Engineering Optimization?

被引:40
作者
Bernardez, Guillermo [1 ]
Suarez-Varela, Jose [1 ]
Lopez, Albert [1 ]
Wu, Bo [2 ]
Xiao, Shihan [2 ]
Cheng, Xiangle [2 ]
Barlet-Ros, Pere [1 ]
Cabellos-Aparicio, Albert [1 ]
机构
[1] Univ Politecn Cataluna, Barcelona Neural Networking Ctr, Barcelona, Spain
[2] Huawei Technol Co Ltd, Network Technol Lab, Beijing, Peoples R China
来源
2021 IEEE 29TH INTERNATIONAL CONFERENCE ON NETWORK PROTOCOLS (ICNP 2021) | 2021年
关键词
Traffic Engineering; Routing Optimization; Multi-Agent Reinforcement Learning; Graph Neural Networks;
D O I
10.1109/ICNP52444.2021.9651930
中图分类号
TP3 [计算技术、计算机技术];
学科分类号
0812 ;
摘要
Traffic Engineering (TE) is a basic building block of the Internet. In this paper, we analyze whether modern Machine Learning (ML) methods are ready to be used for TE optimization. We address this open question through a comparative analysis between the state of the art in ML and the state of the art in TE. To this end, we first present a novel distributed system for TE that leverages the latest advancements in ML. Our system implements a novel architecture that combines Multi-Agent Reinforcement Learning (MARL) and Graph Neural Networks (GNN) to minimize network congestion. In our evaluation, we compare our MARL+GNN system with DEFO, a network optimizer based on Constraint Programming that represents the state of the art in TE. Our experimental results show that the proposed MARL+GNN solution achieves equivalent performance to DEFO in a wide variety of network scenarios including three real-world network topologies. At the same time, we show that MARL+GNN can achieve significant reductions in execution time (from the scale of minutes with DEFO to a few seconds with our solution).
引用
收藏
页数:11
相关论文
共 52 条
[1]  
Almasan P., 2019, ARXIV191007421
[2]  
Andrychowicz M., 2021, INT C LEARN REPR
[3]  
[Anonymous], 1998, OSPF Version 2, RFC 2328, Internet Engineering Task Force
[4]  
[Anonymous], 1999, 2702 RFC
[5]  
[Anonymous], 2012, White Paper
[6]  
[Anonymous], 2018, P WORKSH BIG DAT AN, DOI DOI 10.1145/3229607.3229610
[7]  
Awduche D., 2002, RFC 3272, DOI DOI 10.17487/RFC3272
[8]   Optimal oblivious routing in polynomial time [J].
Azar, Y ;
Cohen, E ;
Fiat, A ;
Kaplan, H ;
Räcke, H .
JOURNAL OF COMPUTER AND SYSTEM SCIENCES, 2004, 69 (03) :383-394
[9]   Stability issues in OSPF routing [J].
Basu, A ;
Riecke, JG .
ACM SIGCOMM COMPUTER COMMUNICATION REVIEW, 2001, 31 (04) :225-236
[10]  
Battaglia PW, 2016, ADV NEUR IN, V29