Smartly Managing Traffic and Latency in a Content-Based data center using Modified Packet Headers, Machine Learning, Load Balancing, and Network Telemetry
Souryendu Das, Andrew White, Malcolm Lyn, Stavros Kalafatis · 2025
This paper introduces SRLBA, an approach for optimizing request management in data centers, specifically addressing the challenges of routing and load balancing within server racks. Our primary objective is to derive efficient routing and load-balancing decisions that mitigate oversubscription, ensuring effective server utilization. SRLBA leverages Software-Defined Networking (SDN) and established Machine Learning (ML) techniques to classify packets and balance traffic. The method integrates SDN-based switches, OpenFlow for packet header modification, and ML algorithms for traffic classification and workload prediction. Unlike approaches relying on reinforcement learning (RL), SRLBA exclusively utilizes supervised learning, ensuring efficiency and scalability within modern data center environments. We demonstrate SRLBA’s effectiveness using custom and public datasets, offering an innovative solution to request management through the integration of SDN and ML technologies.