Programmable infrastructures for secure healthcare

This thesis addresses the critical challenges of secure and efficient medical data sharing, proposing the Enabling Personalized Intervention (EPI) framework to advance personalized medicine and Digital Health Twins (DHTs). With the growing diversity and volume of health data—from electronic health records to genomics and wearable devices—secure collaboration among healthcare providers, researchers, and stakeholders is paramount. The EPI framework provides a comprehensive solution for policy-compliant, dynamic data-sharing workflows, emphasizing privacy and security by design.
The EPI framework integrates data-sharing logic models with formalized policy enforcement, enabling automated reasoning and secure data exchanges across institutions. This thesis develops adaptive Service Function Chain (SFC) provisioning methods, employing heuristic-boosted deep Q-learning to dynamically manage resource allocation. By containerizing network functions, the framework facilitates rapid deployment, high reusability, and low overhead in orchestrating workflows. These features support diverse healthcare use cases, balancing performance, privacy, and compliance.
To address privacy and security risks, the framework incorporates advanced risk assessment models, aligning data-sharing workflows with predefined policies and data utility constraints. Experimental evaluations demonstrate the framework's capacity to mitigate risks while optimizing data movement and access control, fostering trust among stakeholders. By bridging gaps in policy reasoning, network orchestration, and privacy protection, the EPI framework sets a new standard for secure medical data-sharing ecosystems. This work underscores the transformative potential of data-driven healthcare, laying the groundwork for scalable, privacy-aware Digital Health Twin (DHT) applications that can revolutionize patient care and research collaboration.