Extending the Migration from Asynchronous to Reactive Programming in Java: A Performance Analysis of Caching, CPU-Bound, and Blocking Scenarios
Applied Sciences · 21.12.2025 · 16 (1) · 90
DOI: 10.3390/app16010090 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 21.12.2025. Verfasst von Andrei Zbarcea und 2 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Modern distributed systems increasingly rely on reactive programming to meet the demands of high throughput and low latency under extreme concurrency. While the theoretical advantages of non-blocking I/O are well-established, empirical understanding of its behavior across heterogeneous enterprise workloads remains fragmented. This study presents a unified architectural evaluation of asynchronous (thread-per-request) and reactive (event-loop) paradigms within a functionally equivalent Java microservice environment. Unlike prior studies that isolate specific workloads, this research benchmarks the architectural crossover points across three distinct operational categories: distributed caching, CPU-bound processing, and blocking I/O, under loads up to 1000 concurrent users. The results quantify specific boundary conditions: the reactive model demonstrates superior elasticity in I/O-bound caching scenarios, achieving 75% higher throughput and 68% lower memory footprint. However, this advantage is strictly workload-dependent; both paradigms converge to an identical CPU wall at processor saturation, where the reactive model incurs a quantifiable latency penalty due to event-loop contention. Furthermore, under blocking conditions, the reactive model’s memory efficiency (reducing footprint by ~50%) provides resilience against Out-Of-Memory (OOM) failures, even as throughput gains plateau. These findings move beyond generic performance comparisons to provide precise, data-driven guidelines for hybrid architectural adoption in complex distributed systems.
Zitieren
APA
Zbarcea, A., Tudose, C., & Boicea, A. (2025). Extending the Migration from Asynchronous to Reactive Programming in Java: A Performance Analysis of Caching, CPU-Bound, and Blocking Scenarios. Applied Sciences, 16(1), 90. https://doi.org/10.3390/app16010090
Vancouver
Zbarcea A, Tudose C, Boicea A. Extending the Migration from Asynchronous to Reactive Programming in Java: A Performance Analysis of Caching, CPU-Bound, and Blocking Scenarios. Applied Sciences. 2025;16(1):90. doi:10.3390/app16010090
BibTeX
@article{Zbarcea2025,
author = {Zbarcea, Andrei and Tudose, Cătălin and Boicea, Alexandru},
title = {Extending the Migration from Asynchronous to Reactive Programming in Java: A Performance Analysis of Caching, CPU-Bound, and Blocking Scenarios},
journal = {Applied Sciences},
year = {2025},
volume = {16},
number = {1},
pages = {90},
doi = {10.3390/app16010090}
}
EUVIMED