Containerization of User Interface Application - Performance Impact of Containerization

Abstract

Containerization has become a dominant strategy in modern software deployment, offering benefits such as improved portability, security, and environment consistency. Although its advantages in server-side applications are well-established, its impact on graphical user interface (UI) applications remains underexplored, particularly in embedded systems with constrained resources. This thesis investigates the performance implications of containerizing UI applications by conducting empirical tests on a Chromium-based browser operating in both native and containerized environments on an ARM-based embedded Linux platform. The study evaluates performance on key metrics, including CPU usage, memory consumption, JavaScript execution, and rendering throughput. Quantitative analysis revealed that containerization introduces negligible CPU overhead (approximately +0.9%) and has no significant impact on JavaScript execution performance. However, it significantly increases memory usage by 31.7% and reduces rendering performance by 7.84% on average, as measured by standardized MotionMark benchmarks. The findings indicate that while scripting-heavy applications may remain viable within containers, graphics-intensive workloads are more susceptible to container-induced performance degradation. These results suggest that performance trade-offs must be carefully considered when deploying containerized user interface applications in latency sensitive or resource-constrained environments. The study contributes to understanding the practical limitations of containerization and outlines directions for future research, including GPU acceleration and a broader evaluation across UI frameworks and hardware platforms.