MiCADO -Microservice-based Cloud Application-level Dynamic Orchestrator

Various scientific and commercial applications require automated scalability and orchestration on cloud computing resources. However, extending applications with such automated scalability on an individual basis is not feasible. This paper investigates how such automated orchestration can be added to cloud applications without major reengineering of the application code. We suggest a generic architecture for an application level cloud orchestration framework, called MiCADO that supports various application scenarios on multiple heterogeneous federated clouds. Besides the generic architecture description, the paper also presents the first MiCADO reference implementation, and explains how the scalability of the Data Avenue service that is applied for data transfer in WS-PGRADE/gUSE based science gateways, can be improved. Performance evaluation of the implemented scalability based on up and downscaling experiments is presented