REST API Response Time Optimization Using Elasticsearch

This study compares the performance of relational databases and unstructured databases, specifically Elasticsearch, in optimizing REST API response times. As dataset sizes increase, relational databases face prolonged data retrieval times, a common challenge for many companies relying on them for general-purpose use. Using a replica of an online coffee shop's product catalogue dataset, this research analyzes data retrieval scenarios presented in JSON format. The primary goal is to enhance REST API response times by implementing Elasticsearch, leveraging the inherent speed advantage of unstructured databases in data retrieval. The process involves replicating the catalogue data, converting it to relational JSON, setting up a Node.js server with Elasticsearch runtime, and migrating the data. The results show that Elasticsearch significantly outperforms the relational database, with an average response time of 282.6 ms compared to 754.8 ms—a 62.5596% improvement. Load testing with 200 concurrent customers accessing the REST API demonstrates a throughput of 60 requests per minute and an average response time of 930 ms, ensuring a smooth user experience under heavy demand. In conclusion, replacing relational databases with Elasticsearch optimizes data retrieval speed, emphasizing its practicality in enhancing data retrieval processes.