An Execution Controller Enabling Autonomic Features in the Cloud-Native Environment

Software systems have reached a significant level of complexity. Most applications are designed with Cloud Computing in mind, according to DevOps concepts, CI/CD, containers, and microservices. These applications are referred to as Cloud-native applications. Despite the undeniable benefits, such as scalability, high availability, and efficient system deployment, this paradigm presents several challenges. One of them is ensuring that the system is autonomic, namely that the system is able to automatically adjust its behavior in response to changes in state and external factors. In this paper, we opt for the fact that to enable autonomic features in any system, actions must address the whole execution environment. The proposed architecture of an action distinguishes three of its possible states that are: definition, planning, and processing. In each state, actions can be characterized by properties that cover aspects such as prioritization or idempotency restrictions. Then, we propose an architecture of the execution controller - a component processing the actions and responsible for modification of the Cloud-native execution environment in response to occurring anomalies. The proposed solution was evaluated with respect to performance, management capabilities and the impact on autonomy. Hence, showing improvement of dynamic Cloud-native execution environments.