I work for a group of emergency services in Australia
I've recently thought about creating 3 levels of database classification
Level 1 – Multiple Availability Groups and Multiple Servers – Include AG1 for High Security, High Availability AG2 1. Updates, etc. can be applied to servers in this environment in a progressive manner.
Level 2 – Single Failover Cluster – contains non-essential services for production, such as reporting services.
Level 3 – Includes non-production environments and maybe flipping clusters for specific tests
Some things that I would think to be an advantage of this model is
Logging for sensitive datasets can be set up on an individual AV group – according to the wishes of the company. Service Pack tests, etc. can be implemented on the production to get the final UAT while having a withdrawal plan
Migration to the cloud or "as a service" will be less painful.
So my questions are
Would you consider creating different database environments based on security and operational requirements?
Do you agree with this as a strategy?
Is there another option proposed?
How would you sell the concept with management – as this adds to the complexity of management resource requirements and day-to-day implementation?