Comparison of JSON and MessagePack Serialization Performance on End-to-end latency over HTTP Protocol

Muhammad Ilham Dzurinadib (Universitas Nahdlatul Ulama Blitar, Indonesia)
Muhamat Maariful Huda (Universitas Nahdlatul Ulama Blitar, Indonesia)
Tito Prabowo (Universitas Nahdlatul Ulama Blitar, Indonesia)

G-Tech : Jurnal Teknologi Terapan
Vol 10 No 3 (2026): G-Tech, Vol. 10 No. 3 July 2026

The increasing use of HTTP-based communication in modern service-oriented systems requires efficient data serialization formats to reduce communication latency and transmission overhead. This study compares the performance of JSON and MessagePack serialization formats in HTTP communication using the Go programming language. Performance was evaluated using synthetic datasets containing 10, 100, 1000, and 5000 records. The measured parameters included payload size, encoding latency, decoding latency, and End-to-end latency. The results show that MessagePack consistently outperformed JSON across all evaluated metrics. For the largest dataset containing 5000 records, MessagePack reduced End-to-end latency from 13.269 ms to 3.894 ms and generated a smaller payload size (335.87 KB) than JSON (350.53 KB). The performance advantage became more pronounced as the dataset size increased, particularly in decoding and End-to-end latency processes. These findings indicate that MessagePack is an efficient alternative to JSON for HTTP-based communication systems requiring low latency, bandwidth efficiency, and high-performance data exchange.