Peering Partner Recommendation for ISPs using Machine Learning

2025-09-11

AI Summary

This paper addresses the inefficiency and protracted duration of ISP peering decisions by proposing the first machine learning–based automated peering partner recommendation framework. Methodologically, it systematically constructs a heterogeneous feature set integrating multi-source public data—including PeeringDB and CAIDA—and conducts the first comparative evaluation of XGBoost, neural networks, and Transformers for this task. Results demonstrate that XGBoost achieves 98% accuracy—significantly outperforming alternatives—while exhibiting superior generalizability across geographic regions and time periods, robustness to missing data, and computational efficiency during inference. The framework provides ISPs with a production-ready, fully automated decision-support tool for peering. Moreover, it establishes a novel paradigm in Internet governance: leveraging structured infrastructure data with robust tree-based models to inform critical network interconnection policies.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Recommender Systems
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Internet service providers (ISPs) need to connect with other ISPs to provide global connectivity services to their users. To ensure global connectivity, ISPs can either use transit service(s) or establish direct peering relationships between themselves via Internet exchange points (IXPs). Peering offers more room for ISP-specific optimizations and is preferred, but it often involves a lengthy and complex process. Automating peering partner selection can enhance efficiency in the global Internet ecosystem. We explore the use of publicly available data on ISPs to develop a machine learning (ML) model that can predict whether an ISP pair should peer or not. At first, we explore public databases, e.g., PeeringDB, CAIDA, etc., to gather data on ISPs. Then, we evaluate the performance of three broad types of ML models for predicting peering relationships: tree-based, neural network-based, and transformer-based. Among these, we observe that tree-based models achieve the highest accuracy and efficiency in our experiments. The XGBoost model trained with publicly available data showed promising performance, with a 98% accuracy rate in predicting peering partners. In addition, the model demonstrated great resilience to variations in time, space, and missing data. We envision that ISPs can adopt our method to fully automate the peering partner selection process, thus transitioning to a more efficient and optimized Internet ecosystem.

Problem

• Automating ISP peering partner selection using machine learning
• Predicting optimal ISP peering relationships from public data
• Enhancing efficiency in Internet connectivity through automated recommendations

Innovation

• Machine learning predicts ISP peering partnerships
• XGBoost model achieves 98% accuracy rate
• Uses public data from PeeringDB and CAIDA