MAGNNETO: A Graph Neural Network-Based Multi-Agent System for Traffic Engineering

Abstract

Current trends in networking propose the use of Machine Learning (ML) for a wide variety of network optimization tasks. As such, many efforts have been made to produce ML-based solutions for Traffic Engineering (TE), which is a fundamental problem in ISP networks. Nowadays, state-of-the-art TE optimizers rely on traditional optimization techniques, such as Local search, Constraint Programming, or Linear programming. In this paper, we present MAGNNETO, a distributed ML-based framework that leverages Multi-Agent Reinforcement Learning and Graph Neural Networks for distributed TE optimization. MAGNNETO deploys a set of agents across the network that learn and communicate in a distributed fashion via message exchanges between neighboring agents. Particularly, we apply this framework to optimize link weights in OSPF, with the goal of minimizing network congestion. In our evaluation, we compare MAGNNETO against several state-of-the-art TE optimizers in more than 75 topologies (up to 153 nodes and 354 links), including realistic traffic loads. Our experimental results show that, thanks to its distributed nature, MAGNNETO achieves comparable performance to state-of-the-art TE optimizers with significantly lower execution times. Moreover, our ML-based solution demonstrates a strong generalization capability to successfully operate in new networks unseen during training.

Autonomous Flow Routing for Near Real-Time Quality of Service Assurance

“…Many works have proposed solutions for Traffic Engineering (TE) that optimize link weights for minimizing network congestion, e.g., based on heuristics, as in [11]. Targeting at significantly lower execution times, which is of paramount importance in the case of highly dynamic traffic scenarios, other works have proposed ML techniques for TE optimization, like [12] and [13], where the use of DRL was proposed, and [14] that leverages MAS for distributed TE optimization. Multiprotocol Label Switching (MPLS) is another routing technology that can be combined with TE for traffic steering [16].…”

Multi-agent graph learning-based optimization and its applications to computer networks

“…Therefore, we claim that MAGNNETO is among the first architectures that embeds a GNN into a MARL setting to provide agents with topology-awareness and generalization capabilities while simplifying the cooperation mechanisms, and the first one that addresses the deployability challenges in real-world networked contexts (Internet Service Provider (ISP) Networks [32,33], DCNs [34], Power Grids [58]). In next Section 3.3 we provide further details about the main contributions of MAGNNETO.…”

“…In our second work about MAGNNETO-TE [33], we deep dived into the question of our first paper by formulating an enhanced MAGNNETO framework (the final version that we described in Chapter 3) and performing a much more comprehensive evaluation. We summarize below the main differences of the extended evaluation presented in this Section with respect to the previous exploratory one:…”