Improving cloud architectures using UML profiles and M2T transformation techniques

Abstract

In this paper, we present an approach with the goal to improve the underlying architecture of cloud systems. For this, we propose UML2Cloud, a framework targeted at modeling and checking cloud systems. The main core of UML2Cloud uses UML profiles to capture the main elements of a cloud system including, among other elements, its underlying architecture and the interaction with clients. Additionally, UML2Cloud uses Model-to-Text transformation techniques to automatically generate configuration documents representing complex cloud scenarios. In this work, we use these documents as input for a cloud simulation tool, called Simcan2Cloud, to simulate the behavior of different systems. Thus, the analysis of the performance results obtained from the simulations allows us to draw some conclusions about how to improve the efficiency of the studied clouds by adjusting the hardware resource configuration.