Container-based Architecture to Optimize the Integration of Microservices into Cloud-Based Data-Intensive Application Scenarios

Earth Observation data archives are currently growing at unprecedented speeds. New satellites add petabytes of data every year. At the same time, the amount of data provided by Earth-based in-situ networks is growing at enormous rates. These developments have led to a change in data processing paradigms. Data is not down-loaded and processed locally anymore, but applications are sent to the data. This paper demonstrates an Big Data architecture that allows for interoperable solutions across data providers, integrators, and users. The availability of a mature domain architecture as provided by the Open Geospatial Consortium provides a solid base and allows for uniform microservice handling. The makro- and microarchitecture described herein uses self-contained Docker images to allow for transparent microservices, horizontal scale-out, and high reliability and maintainability thanks to decoupled and self-sustained execution elements.