Clouds reduces time-to-market and provides on-demand scalability at a low cost. Many private and public Clouds have emerged during the last years, offering a range of different technologies each suited for particular types of applications. SeaClouds tackles the problem of deploying and managing, in an efficient and adaptive way, complex multi-services applications over technologically heterogeneous Clouds. This allows organisations to embrace Cloud solutions and avoid risks of unreliability and vendor lock-inSeaClouds provides an answer to questions such as:- How can complex applications be deployed, managed and monitored over multiple Clouds, distributing its modules according to the deployment requirements and the strong and weak points of each offering?- How can complex applicatiosn be reconfigured if run-time problems are detected?SeaClouds' approach is based on the concept of cloud-based services orchestration. Orchestration means the automated arrangement, coordination and management of multiple services as a single aggregated complex application, without the need to modify the code of the services. Orchestration can be formally designed to fulfil properties such as QoS of the whole application. Application reconfiguration is also performed by changing the orchestration of the services, when the monitoring detects that certain properties such as QoS are not respected.The objectives of SeaClouds are therefore:- Orchestration, adaptation, and verification of services distributed over different Clouds- Monitoring and runtime reconfiguration of services distributed over different Clouds- Provision of a unified Cloud-independent mechanism to manage services distributed over different Clouds.- Alignment of SeaClouds's architecture with major standards for cloud interoperability, particularly OASIS' CAMP and TOSCA, promoting them in research and industrial communitiesSeaClouds aims to support developers and application managers; for these target groups SeaClouds will provide:- In the design phase:o\tA new approach based on TOSCA to express how each component of the application should interact with the otherso\tA language to specify requirements in terms of QoS for each components and for the application as a whole- In the deployment phase:o\tTool for searching among existing Clouds in order to find those that best match the developer's requirements expressed at design timeo\tTools to deploy the application on the selected Clouds- In the runtime phase:o\tTools to monitor and analyse the performances of each components across different providerso\tTools to assess whether and which components should be redeployed on different Cloud providers, in case of non-satisfactory performanceso\ttools to redeploy the underperforming components on different Clouds and to adapt the orchestration to the new configurationSuch tools will be organized in a framework, which will be available either as software to install on premises or as SaaS, or both