IntelliStore: An Intelligent AI Agent Framework for Autonomous Storage and Database Optimization in Cloud-Native Microservices

Abstract

Cloud-native microservices architectures face significant challenges in optimizing storage and database configurations across diverse, dynamically scaling services. Traditional approaches require manual intervention and service-specific tuning, leading to suboptimal resource utilization and increased operational costs. This paper presents IntelliStore, a novel intelligent agent framework that autonomously identifies optimal storage technologies and database configurations for microservice applications in cloud environments. Our system employs a multi-agent architecture that continuously monitors storage usage patterns, analyzes workload characteristics, benchmarks against published performance metrics, and generates actionable recommendations for both deployed and prospective services. IntelliStore leverages large language models for intelligent decision-making, combining real-time metrics collection with historical performance data to suggest optimal storage backends, database types, and configuration parameters. We evaluate our system on production microservices handling varying workloads, demonstrating an average 34% reduction in storage costs, 28% improvement in I/O performance, and 42% decrease in configuration tuning time compared to manual optimization approaches. Our results show that AI-driven autonomous storage optimization can significantly enhance resource efficiency while maintaining service-level agreements in large-scale cloud deployments.