The widespread adoption of Free and Open Source Software (FOSS) in many strategic contexts of the information technology society has drawn the attention on the issues regarding how to handle the complexity of assembling and managing a huge number of (packaged) components in a consistent and effective way. FOSS distributions (and in particular GNU/Linux-based ones) have always provided tools for managing the tasks of installing, removing and upgrading the (packaged) components they were made of. While these tools provide a (not always effective) way to handle these tasks on the client side, there is still a lack of tools that could help the distribution editors to maintain, on the server side, large and high-quality distributions. In this paper we present our research whose main goal is to fill this gap: we show our approach, the tools we have developed and their application with experimental results. Our contribution provides an effective and automatic way to support distribution editors in handling those issues that were, until now, mostly addressed using ad-hoc tools and manual techniques.

We introduce the Aeolus component model, which is specifically designed to capture realistic scenarii arising when configuring and deploying distributed applications in the so-called cloud environments, where interconnected components can be deployed on clusters of heterogeneous virtual machines, which can be in turn created, destroyed, and connected on-the-fly. The full Aeolus model is able to describe several component characteristics such as dependencies, conflicts, non-functional requirements (replication requests and load limits), as well as the fact that component interfaces to the world might vary depending on the internal component state.When the number of components needed to build an application grows, it becomes important to be able to automate activities such as deployment and reconfiguration. This correspond, at the level of the model, to the ability to decide whether a desired target system configuration is reachable, which we call the achievability problem, and producing a path to reach it.In this work we show that the achievability problem is undecidable for the full Aeolus model, a strong limiting result for automated configuration in the cloud. We also show that the problem becomes decidable, but Ackermannhard, as soon as one drops non-functional requirements. Finally, we provide a polynomial time algorithm for the further restriction of the model where support for inter-component conflicts is also removed.Keywords: software component, model, cloud computing, distributed systems ✩ This paper is an extended and revised version of [1, 2] -a detailed comparison with these papers is reported in the related work Section 5. This work has been developed within the ANR-2010-SEGI-013-01 project Aeolus "Mastering the Complexity of the Cloud": we would like to thank all the project partners for the numerous discussions that contributed to the definition of the Aeolus model. This work has been partially performed at IRILL, center for Free Software Research and Innovation in Paris, France,