Dynamic Adaptation for Distributed Systems in Model-Driven Engineering

Modern-day software systems operate within complex, uncertain, and highly dynamic environments. Managing such systems is a significant challenge; developing self-managing autonomic systems is one way to reduce development and maintenance efforts. In the context of distributed systems, achieving this autonomy through dynamic adaptation is particularly challenging due to the volatile host environment. Model-Driven Engineering (MDE) is a software development paradigm that advocates the use of models as the primary artifacts rather than source code. MDE promises higher-quality software at a lower cost through abstractions, automation, and analyses. The goal of our work is to leverage MDE to facilitate the development and maintenance of distributed applications with dynamic adaptation capabilities. We assume that the structure and behavior of the application has been modeled using the Componentand-Connector (C&C) paradigm and the 'Monitor-Analyze-Plan-Execute with shared Knowledge' (MAPE-K) reference architecture. In the initial work, we have developed a model-level monitoring infrastructure, and adapted existing code generation and deployment support to generate a distributed system from the C&C models and deploy it automatically on a suitable platform. In future work, we plan to investigate how the monitoring and adaptation capabilities of cloud-native containerization and orchestration platforms (i.e., Docker and Kubernetes) can be leveraged for dynamic adaptation, and how this system-level adaptation can be combined effectively with any model-level monitoring, planning, and adaptation capabilities.