ATMOSPHERE (Adaptive, Trustworthy, Manageable, Orchestrated, Secure Privacy-assuring Hybrid, Ecosystem for REsilient Cloud Computing) is a 24-month project aiming at the design and development of an ecosystem, of a framework, a platform and applications of next-generation trustworthy cloud services on top of an intercontinental hybrid and federated resource pool. The framework considers a broad spectrum of properties and their measures. The platform supports the building, deployment, measurement and evolution of trustworthy cloud resources, data network and data services. The platform is demonstrated in a sensitive scenario to build a cloud-enabled secure and trustworthy application related to distributed telemedicine.

This report belongs to the WP4, “Cloud services managing Federated and Hybrid resources”. The objective of this work package is to develop an orchestration platform and a dashboard to manage a federated set of container-based resources, as well as to provide measures, adaptive mechanisms and policies to improve trustworthiness of data management and processing services. The platform will provide federation at two levels: at the level of cloud resources and at the level of containers. Several aspects need to be considered at both levels: authentication and authorisation, container image management, and networking. Pooling at the level of cloud resources will leverage the Fogbow middleware. The platform will also be capable of managing hybrid resources such as specialized devices (e.g., GPGPUs) and will facilitate the development of serverless applications.
The first version of the platform includes the capability of the federation of cloud resources from sites located in different continents, the services to deploy applications, the inventory of services and repositories that are needed to interact with the federated infrastructure and several examples of distributed elastic applications. It also includes the ability to connect resources instantiated in different cloud providers through a single federated private network.

The architecture of the system is in continuous refinement since its first release in March of this year. It has been refined to describe the way federated networks will be implemented, the integration of the different layers and the way resources are managed. The sample applications are extracted from the definitions given in D7.2 Design of the Biomarker Pilot Application, including different complexity levels. The next version of the infrastructure will include the support to heterogeneous resources and specific hardware, although the current application architecture are defined in a way to support it.