International Journal of Scientific Research and Engineering Development

International Journal of Scientific Research and Engineering Development


( International Peer Reviewed Open Access Journal ) ISSN [ Online ] : 2581 - 7175
Submit Your Manuscript OnlineIJSRED

Proactive ML-Based Load Balancing for Cloud Microservices


📑 Paper Information
📑 Paper Title Proactive ML-Based Load Balancing for Cloud Microservices
👤 Authors Prof. Roopa B, Madhusudhan U, Rakesh D, Goutham R, Kiran Kumar
📘 Published Issue Volume 9 Issue 6
📅 Year of Publication 2026
🆔 Unique Identification Number IJSRED-V9I6P61
📝 Abstract
Cloud-based applications developed using microservices architecture experience continuous variations in workload due to changing user requests. During peak traffic conditions, increased resource utilization can cause server overload, higher response time and reduced application performance. Traditional load-balancing and autoscaling techniques generally follow a reactive approach, where additional resources are allocated only after predefined resource-utilization thresholds are reached. This may result in delayed responses during sudden increases in workload. To address these challenges, this paper presents a Proactive Machine Learning-Based Load Balancing framework for cloud microservices. The proposed system uses a Long Short-Term Memory (LSTM) network to analyse historical workload information and predict future CPU utilization. The prediction process considers five important parameters: current users, CPU utilization, memory usage, requests per second and response time. A rule-based mechanism converts the predicted CPU utilization into recommended backend pod counts to support proactive resource management. The system is developed using Python and Fast API for backend services and React for the frontend interface. A workload simulator generates different traffic conditions, while Prometheus and Grafana provide resource monitoring and visualization. Docker and Kubernetes support containerization and deployment, with Horizontal Pod Autoscaler (HPA) configuration for replica management. The framework integrates workload generation, metric collection, machine learning-based prediction and container orchestration to support improved scalability and resource utilization in cloud-hosted applications.
📝 How to Cite
Prof. Roopa B, Madhusudhan U, Rakesh D, Goutham R, Kiran Kumar, "Proactive ML-Based Load Balancing for Cloud Microservices" International Journal of Scientific Research and Engineering Development, V9(6): Page(445-452) November - December 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.