Software System Adaptability: A Middleware for Orchestrating Legacy Component Behavior

Abstract

The increasing need for software systems to adapt dynamically to changing operational contexts has highlighted the challenge of integrating legacy components, which were often designed without adaptability in mind. This paper introduces Tekio, a lightweight middleware framework built on the OSGi specification that enables legacy software libraries to function effectively within self-adaptive systems. Tekio incorporates the MAPE-K feedback loop to support dynamic reconfiguration through parameter adjustment, component replacement, context-event response, and quality-of-service adaptation. A case study in computer vision using the OpenCV library demonstrates that Tekio achieves near-native performance while enabling frequent runtime adaptations. Empirical evaluation shows minimal overhead in throughput, CPU utilization, and memory consumption, with settling times in the millisecond range for low-to-medium complexity adaptations. The results indicate that Tekio provides a viable pathway for extending the operational lifespan of legacy software in modern adaptive environments, balancing flexibility with performance efficiency.