Observability in Data Spaces

Data spaces are meant to be a trusted system where participants can share, and exchange data under a controlled governance scheme. Even though data sharing happens always between the participants under some circumstances, such transactions need to be observed for different reasons. Observability of the transaction can be a requirement only between the parties involved or as a measure indicated by regulations or the governance scheme of the data space. This document shall investigate the need for observability in data spaces, reflect the current findings on the subject, provide some insights into relevant use cases and propose potential solutions.

The need for observability and requirements is stated in the IDSA Rulebook. Observability may result in the exchange of regulated or highly valuable datasets. Likewise, marketplaces or general billing and charging may require the observation of the transaction. Depending on the requirements, this could lead to a centralized or a decentralized solution to realize the observability of the transaction. The approach to identifying an appropriate solution depends on the requirements stated in the general governance framework of the data space, which is managed by the Data Space Governance Authority or the individual contract between the participants. Furthermore, it is important to consider for whom the collected information will serve after it is collected. Different kinds of third parties may need to be enabled to read or analyze the collected information, e.g., a third-party auditor, a governmental authority, or any third party identified in case of incidents. Such information needs to be provided or evaluated before the transaction, during the transaction, or even after the transaction.

In this regard, observability in data spaces enhances trust by attesting that data is used as agreed. It aligns with data spaces' governance principles and strengthens compliance monitoring and, therefore, protects data providers' and consumers’ interests. Observability imposes some technical complexity of tracking data usage and has privacy implications for monitoring. Legal constraints across jurisdictions, potential resource-intensive implementation, and subsequently a potential impact on data processing performance need to be addressed.