Design and Performance Evaluation of Event-Driven Microservices for Large-Scale Enterprise Data Processing

Modern enterprise systems require scalable, reliable, and real-time data processing capabilities, which traditional monolithic architectures cannot efficiently support under high-load conditions. This research proposes an Event-Driven Microservices Architecture (EDMA) integrated with cloud-native technologies, including Apache Kafka, Kubernetes, observability tools, DevSecOps practices, and AI-assisted orchestration. The study adopts an experimental and quantitative approach using synthetic enterprise workloads, fault-injection scenarios, and comparative analysis with monolithic and traditional microservices systems. Performance evaluation is conducted using key metrics such as throughput, latency, scalability, reliability, and resource utilization. The results demonstrate that the proposed EDMA significantly improves system performance by increasing throughput, reducing latency, enhancing scalability, and strengthening fault tolerance while optimizing computational resources. Furthermore, AI-driven orchestration and integrated observability mechanisms enable adaptive resource management, predictive optimization, and improved operational efficiency under dynamic workloads. Overall, the findings confirm that combining event-driven design with cloud-native infrastructure provides a robust, efficient, and scalable solution for modern large-scale enterprise data processing systems.