What does best practice architecture look like for your Azure Data Lake in your organisation? How have you implemented RBAC/ACL to ensure data is made available only to those required to consume?

258 viewscircle icon1 Comment
Sort by:
Director of Data2 years ago

I'm working on something like this right now. What we recommend often is the kind of gold-silver-bronze layer approach of the data lake. Bronze is your raw data, gold is the end consumer and silver is some kind of hybrid of the two. And yes, we have designed our backend RBAC/ACL's for each of those layers. So within each layer, the data organized in a taxonomy that is tagged to particular security access for individual user groups. The gold layer has broader accessibility because it's been fully managed and secured. Silver and bronze have a more specific user group because they need more training and more sophistication in that using the tools. But again, it's organized by subject area or data domain. And this for this particular client has to be split between regulatory and non-regulatory - and certain subsets of that data set more tightly controlled. So in that RBAC model, we've got this kind of split based on data lake area, domain and data classification.

Lightbulb on1

Content you might like

Strongly agree9%

Agree52%

Neutral16%

Disagree20%

Strongly disagree1%

Other (please specify)

View Results

Short-sighted focus on profit18%

Lack of holistic view of data pipelines42%

Regulations not keeping up19%

Unclear ownership of responsibility for ethical data use16%

Limited view of potential impact of unethical data use5%

Something else? Comment below

View Results