Towards modelling support for multi-cloud and multi-data store applications

Abstract

The support to cloud enabled databases varies from one cloud provider to another. Developers face the task of supporting applications living in different clouds, and therefore of supporting different database management systems. To these developers, the challenge lies in understanding the differences in expressivity between data stores and the impact of such differences on the rest of the application. The advent of the NoSQL movement increased the complexity of this task by leveraging the creation of a large number of cloud enabled database management systems employing slightly different data models. In this paper, we will present a modelling approach that will allow developers to consider the impact of these features to different concrete deployment scenarios in multiple clouds. This approach is currently being developed on the JUNIPER and MODAClouds FP7 projects.