ATMOSPHERE (Adaptive, Trustworthy, Manageable, Orchestrated, Secure Privacy-assuringHybrid, Ecosystem for REsilient Cloud Computing) is a 24-month project aiming at the designand development of an ecosystem of a framework, platform and application of nextgeneration trustworthy cloud services on top of an intercontinental hybrid and federatedresource pool.

The framework considers a broad spectrum of properties and their measures. The platformsupports the building, deployment, measuring and evolution of trustworthy cloud resources,data network and data services. The platform is demonstrated on a sensitive scenario to builda cloud-enabled secure and trustworthy healthcare application that uses machine learningtechniques to aid physicians in their activity of disease diagnosing.

This deliverable is the final result of Task 3.3 Trustworthiness Measurement and AnalysisServices under WP3. The key objective of WP3: Trustworthiness Monitoring & AssessmentFramework is to design and implement the trustworthiness monitoring and assessmentplatform and solutions for the ATMOSPHERE ecosystem. This includes a monitoring platform, atrustworthiness assessment framework, measurement instruments, analysis services, andadaptation capabilities. In particular, this deliverable - D3.6 Trustworthiness Measurementand Analysis Services - is focused on the implementation of the following solutions taken bythe WP3: the Quality Models composition and registration, the probes and actuators providedby each layer of the ATMOSPHERE ecosystem to collect data from the monitored system, the Trustworthiness Score and Metrics computation, the Dashboard service and the interfacesthrough which all the components interact with.

Four quality models are included: i) Cloud Federated Resources (to characterize thetrustworthiness of the federated resources layer, namely services, resources and networklinks, and focusing on availability, isolation, performance and latency aspects); ii) Cloud &Container Services Management (based on the availability, performance, scalability andisolation of virtual services and virtual resources); iii) Data Management Layer (consideringsecurity, performance and privacy of Vallum, and confidentiality, performance and faulttolerance of database engines); and iv) Data Processing Services (to characterize fairness,stability, transparency and privacy of workflows and machine learning models).

Probes and actuators for supporting the computation of the Quality Models are presented,including probes for privacy monitoring at the data storage layer, CLUES, Fogbow, SoftwareDefined Networks, and Kubernetes. As for actuators, the deliverable presents actuators forpreventing policy privacy violations and Kubernetes scaling.

The TMA (Trustworthiness Monitoring & Assessment) framework and the componentsdescribed in this deliverable are open source and freely available at: .