Share
Export Citation
Performance Analysis of Kotlin Multiplatform on Android in a Model-View-Intent (MVI) Architecture Pattern
Afina I.F.
International Conference on Computer and Applications Icca 2025 Proceedings
Abstract
Cross-platform development has gained momentum as mobile applications increasingly demand faster development cycles without compromising runtime performance. Kotlin Multiplatform (KMP) addresses this need by enabling shared business logic while preserving near-native runtime performance, whereas the Model-View-Intent (MVI) architecture promotes predictable state management through unidirectional data flow. This study empirically evaluates KMP with MVI against native Android and Flutter implementations. Results show that KMP achieves startup performance close to native Android (493.6 ms vs. 435.9 ms) and significantly faster than Flutter (1129.2 ms). Memory usage measured using Total Proportional Set Size (PSS) indicates that KMP remains comparable to native Android, whereas Flutter exhibits substantially higher consumption (257.1 MB). One-way ANOVA reveals statistically significant platform effects across all evaluated runtime metrics, with large to very large effect sizes (η² = 0.83–0.99). These findings highlight the trade-offs between code sharing, architectural overhead, and runtime performance in modern cross-platform development.