Protocol for Trustful Data Consumption from Data Repositories by Workflow Engines

Data repositories spread across regions and processed by cloud workflows require fine-grained trust models to ensure that data sovereignty and access restrictions are met. Interactions between the data owner, consumer, repository, workflow and the workflow engine need to managed. We present a three-token protocol designed to ensure secure and trusted execution of workflows that access data from federated repositories, and implement this within the IUDX repository when accessed by a trusted XFaaS workflow engine. Our results indicate modest overheads for the protocol while enforcing access and usage policies set by the data owner during workflow execution. This can help explore more robust mechanisms for data processing.