Abstract-In the landscape of cloud computing, the competition between providers has led to an ever growing number of cloud solutions offered to consumers. The ability to run and manage multi-cloud systems (i.e., applications on multiple clouds) allows exploiting the peculiarities of each cloud solution and hence optimising the performance, availability, and cost of the applications. However, these cloud solutions are typically heterogeneous and the provided features are often incompatible. This diversity hinders the proper exploitation of the full potential of cloud computing, since it prevents interoperability and promotes vendor lock-in, as well as it increases the complexity of development and administration of multi-cloud systems. This problem needs to be addressed promptly. In this paper, we provide a classification of the state-of-the-art of cloud solutions, and argue for the need for model-driven engineering techniques and methods facilitating the specification of provisioning, deployment, monitoring, and adaptation concerns of multi-cloud systems at design-time and their enactment at run-time.

No abstract

Dynamically adaptive systems (DAS) enable the continuous design and adaptation of complex software systems, but their main focus is limited to the application itself rather than the underlying platform and infrastructure. Cloud computing, in contrast, enables the management of the complete software stack, but it lacks integration with software engineering approaches, techniques, and methods from DAS. Model-based approaches have been successfully adopted for modelling DAS at design-time and facilitate their adaptation at run-time. Therefore, a natural next step is to adopt model-based approaches to enable cloud-based DAS. In this paper, we present the Cloud Modelling Framework (CLOUDMF), a model-based framework that addresses this issue. It consists of (i) a tool-supported domain-specific modelling language to model the provisioning and deployment of multi-cloud systems, and (ii) a models@run-time environment for enacting the provisioning, deployment and adaptation of these systems.

Runtime software architectures (RSA) are architecture-level, dynamic representations of running software systems, which help monitor and adapt the systems at a high abstraction level. The key issue to support RSA is to maintain the causal connection between the architecture and the system, ensuring that the architecture represents the current system, and the modifications on the architecture cause proper system changes. The main challenge here is the abstraction gap between the architecture and the system. In this paper, we investigate the synchronization mechanism between architecture configurations and system states for maintaining the causal connections. We identify four required properties for such synchronization, and provide a generic solution satisfying these properties. Specifically, we utilize bidirectional transformation to bridge the abstraction gap between architecture and system, and design an algorithm based on it, which addresses issues such as conflicts between architecture and system changes, and exceptions of system manipulations. We provide a generative tool-set that helps developers implement this approach on a wide class of systems. We have successfully applied our approach on JOnAS JEE system to support it with C2-styled runtime software architecture, as well as some other cases between practical systems and typical architecture models.

Abstract. The key point to leverage model-based techniques on runtime system management is to ensure the correct synchronization between the running system and its model-based view. In this paper, we present a generative approach, and the supporting tool, to make systematic the development of synchronization engines between running systems and models. We require developers to specify "what kinds of elements to manage" as a MOF meta-model and "how to manipulate those elements using the system's management API" as a so-called access model. From these two specifications, our SM@RT tool automatically generates the synchronization engine to reflect the running system as a MOF-compliant model. We have applied this approach on several practical systems, including the JOnAS JEE server.