Integrating Graph Databases into the Data Layer of Three-Tier Architecture Applications

Current research on database integration in enterprise systems primarily focuses on single database implementations, with limited exploration of graph database integration into existing architectures. While traditional Relational Database Management Systems (RDBMS) excel in structured data management, they often struggle with complex relationship traversals in modern Enterprise Resource Planning (ERP) systems. This research presents a systematic approach for enhanced database implementation in three-tier architecture applications, specifically addressing the challenges of managing interconnected data structures in ERP systems. The study implements parallel database configurations using MySQL for RDBMS and Neo4j for graph database, evaluating their performance through comprehensive metrics including query execution time, resource utilization, and API performance. Through extensive testing with datasets ranging from one thousand to ten million records, the experimental results demonstrate that RDBMS excels in complex analytical queries with linear scaling up to one million records, while Neo4j shows superior performance in hierarchical data traversal, maintaining consistent performance (1.7350 seconds) even at ten million records. The proposed dual database approach achieves improved system reliability, with Neo4j maintaining 100 percent success rates across all batch sizes compared to RDBMS 90 percent for larger batches, while demonstrating better resource efficiency (1.70 to 2.89 percent versus 2.70 to 3.39 percent). This research contributes to the field by introducing a Schema Complexity Index for quantifying database complexity, providing empirical evidence for optimal database selection in specific use cases, and establishing a framework for dual database integration in three-tier architectures.