Data lake, data mesh, data fabric…No matter what the architecture, you still need data governance. Governance is the roles/responsibilities, practices and policies that help to bring “balance to the force” sought of speak. Governance is that all encompassing layer that acts as an umbrella over whatever data architecture is planned and designed. It helps to organize, bring controls by providing the how (aka practices) to maintain high quality pipelines.
The Data Management Association (DAMA) gives a great visual in the DAM wheel that wraps data governance activities around the entire data management realm. Note that the foundational activities have governance running through – metadata management, data protection and data quality management, which are needed in any architecture.
So, under each of the architectures below, governance needs to be established in the organization to create some guardrails around your data supply chain.
Data lake – dump it all in there; governance can help to avoid building a data swamp by defining lifecycle policies to retire unused data assets.
Data mesh – data by domain; governance can help establish stewardship to define source of truth in each domain.
Data fabric – intelligence to manage various pipelines and hybrid environments resulting in a more holistic, data-centric outcome. Governance can provide controls around creating and maintaining data quality for data pipelines.
Resources
DAMA wheel – https://www.dama.org/cpages/dmbok-2-wheel-images
HEATH Information 