Fast Replica of Polyglot Persistence in Microservice Architectures for Fog Computing.

Aiming to break the software monoliths that traditional approaches usually produce as artifacts, solutions that are based on microservices consist of heterogeneous and independent software platforms to manage applications and data. In this scenario, the term polyglot persistence has been introduced to characterize software solutions where the involved microservices rely on different data storage technologies. Especially in Fog Computing where data are expected to efficiently flow among nodes - usually from the edge to the cloud - the polyglot persistence could have a negative impact since a combination of data replication and transformation is required. The goal of this paper is to study the challenges in data management in Fog Computing when microservices are adopted, and to present a solution which combines the advantages of the physical copy approach performed by network file systems to provide a fast data movement and the ability of the logical copy approach to transform the data. The resulting mix is demonstrated to reduce the time of creating the replica up to 70%.