Dev, Test, Prod continued discussion (Overall DMZ, DLZ, and DP)

[mike-leuer]

We are currently trying lay the ground work on how our Cloud Scale Analytics will work in AZURE.

I have done a few DEVOPS pipeline as applications but this at a bigger scale with a data-lake with lots of pieces. I am used to having a 1:1 compare between test and prod; however, I feel like that is waist of resources/money. I have some questions that will give me a better insight. I really want a Dev and Prod but currently a Production would be enough to get us off the ground. A reference diagram of how dev, test, prod would do wonders in helping me picture it. (Some where here would be good)
https://docs.microsoft.com/en-us/azure/cloud-adoption-framework/scenarios/cloud-scale-analytics/manage-provision-platform#summary

1. General
   I assume Dev, Test, Prod are on separate subscriptions.
   Should tear down Dev, Test when not being used?
   Are the Data Engineers, Product Engineers, Analyst - working on Dev or in Prod to develop new products or land new data?
   I could approach each project as there own DEVOPS practices for each project data landing and data product. I think it would work great with data products but not landing?

2. Data
   How do you deal with data duplication between Dev, Test, and Prod? In theory I would want 1:1 but that does not sound right
   I see some development resources, is this designed to development?
   https://docs.microsoft.com/en-us/azure/cloud-adoption-framework/scenarios/cloud-scale-analytics/architectures/data-landing-zone#data-lake-services

3. Network
   Can the network be smaller on Dev than on Production? (CIDR Range)
   Do you peer networks for dev, test, prod so that you can transfer data on the Microsoft backbone?