Performance Optimization of GraphQL API Through Advanced Object Deduplication Techniques: A Comprehensive Study
Budi Santosa, Awang Hendrianto Pratomo, Midi Wardana Riski, Shoffan Saifullah, Novrido Charibaldi · Journal of Computing Science and Engineering · 2023
This research paper presents a comprehensive analysis of the performance enhancement achieved in GraphQL application programming interfaces (APIs) when using meticulous object deduplication implementation. By integrating advanced techniques into the GraphQL response mechanism, the data size exchanged between servers and clients can be significantly reduced. Rigorous testing against untreated and HTTP-compressed data validates the obtained results, highlighting the presence of substantial improvements across various performance metrics. The applied object deduplication method demonstrates gains in throughput, with a 0.33% increase observed in a 100-page test. Notably, response time analysis reveals enhancements of 11.04% (10 pages), 62.53% (20 pages), and an impressive 95.22% (100 pages). Meanwhile, parsing time evaluation showcases remarkable increases of 75.78% (10 pages), 276.38% (50 pages), and an even more exceptional 309.35% (100 pages). Comparative analysis against HTTP compression further validates the superiority of object deduplication in parsing time efficiency, demonstrating gains of 64.61% (10 pages), 193.76% (50 pages), and 218.07% (100 pages). While the throughput performance remains comparable, slight differences can be observed in response time, with a 0.66% increase (10 pages), a minor decrease of 0.12 (50 pages), and a modest decline of 1.45% (100 pages). This study fills in existing research gaps and provides empirical evidence of the benefits of object deduplication in enhancing GraphQL API performance, thus enabling the effective optimization of GraphQL APIs.