Empowering Computational Governance - Revolutionize the way we govern data

Nugget by Nora Skjelstad

Many organizations are moving towards a decentralized, domain-oriented data management approach, where data is delivered as well-defined products through a shared, self-serve data platform. This concept is known as Data Mesh. However, one might argue that the most challenging pillar of Data Mesh to understand is “Federated Computational Governance.” What does computational governance entail? And how can we balance the autonomy of local domains with the need for common standards to ensure collaboration across domains?

Federation involves balancing common standards with local adaptations and variations in individual data products. Federated computational governance can be viewed from two perspectives. Firstly, we need to establish common standards that data consumers require to ensure predictability and collaboration across the organization, without restricting the autonomy of individual domains. This is where traditional data management practices come into play. For instance, it is essential to determine common standards for representing data models or schemas, selecting appropriate formats, implementing security policies and measures, and, in some organizations, setting guardrails for naming conventions in tabular data. This could also be requirement set out by external regulations.

Secondly, these common standards form the basis of policies and standards as code, which operationalize computational governance. By implementing these standards as requirements for processing, using, and managing data on a self-serve platform with automated tests, the organization ensures that its data products are compliant with both internal policies and regulatory requirements. Additionally, automated monitoring should be implemented to alert the organization if and when breaches occur.

While computational governance is a valuable concept for reducing the time and resources spent on manual processes and decreasing cognitive load within an organization, it remains one of the more challenging pillars of Data Mesh to implement. The algorithms designed to support this automated governance often become complex due to the need to handle large volumes of diverse data types. Additionally, many organizations struggle to ensure interoperability between data products across different domains, which complicates automated governance routines even further.

So, how can these challenges be addressed? On September 17th, Data Management Association Norway (DAMA) and Agile Lab will host a physical meetup at Rebel in Oslo to discuss computational governance in more detail and explore practical implementation strategies. Don’t miss out!

Meld deg p? her: Computational Governance - en mulig tvangstr?ye eller den “nye vinen”

?