Cascade systems as an implementation of a gray-box architecture: a case study in traceability for microservice scaling

Abstract

[EN] There is great potential in leveraging Artificial Intelligence (AI) systems to optimize complex infrastructures, automate difficult tasks, or support autonomy and coordination between networked devices. However, advances in state-of-theart AI often neglect features and/or requirements that businesses care deeply about, namely traceability and explainability. A majority of available research has not explored much the deployment of semi-physical architectures combining fuzzy rulebased systems with more opaque models to improve explainability, being this specially true for the management of microservices in cloud and cloud-edge environments. This contribution builds on previous work that proposes a middle ground of mixed AI architectures that combine the performance of black-box AI models with ¿a more explainable overall architecture¿ by implementing a microservice scaling system for distributed cloud environments using a cascade approach. This work demonstrates and evaluates an application case of such an approach departing from a Service Level Agreement compliance, in a case of microservice scaling decision over cloud (and cloud-like) infrastructures.