A curious and inquisitive team culture with management support. 
A relational database or similar technology that supports views and joins. 
Trusted identity of the end user at the data layer.

Starting early is best because you can architect the solution into your entire architecture:  data model, database layer, software layer, reporting tools, etc. 

No, infoSecur can be added on to most any data platform.  Depending on your data model, there may be some amount of re-engineering and re-modeling, but usually that is minimal. 

First, there are too many of them.  Second, requirements change and you need to react to change easily.  Third, when you combine users, their roles, and the data, the permutations of security requirements grow exponentially and become unmanageable.

You could, but that is not consistent.  You are trusting that a team or teams of developers, reporting analysts, and users all implement the security rules in the same way and understand them sufficiently to be consistent and thorough.  That is a dangerous assumption. 

First, it’s risky and untrustworthy.  If humans are involved in building those data files, mistakes will be made and data will be inadvertently leaked. It’s just a matter of time. 

Second, it’s not scalable.  You would hope your data becomes more valuable and demands increase over time.  As that progresses, the requirements become complex and the number of data sets (files) grows and the risk grows exponentially with each copy of data you make. 

Third, it wastes everyone’s time – both inside your organization and for your customers/users/partners. 

Since it leverages the database technologies, which today are highly capable of query optimization, it seldom adds noticeable overhead.  At worst, the overhead increase is usually within 5% linearly. Often the query performance actually improves because of the optimization inherent in the methodology.

Data Vault, Dimensional/Star-Schema, and 3rd Normal Form (3NF) are all optimal. Other models like tabular, graph, etc. can be made to accommodate, as can hyper-normalized models like Anchor or Focal.

Yes

Typically Yes

• Health Care HIPAA, Patient Privacy 
• Global Manufacturing  
• Insurance   
• Education
• Government 
• Services 
• Hi Tech
• Trade Secrets 
• Contractual Obligations 
• FERPA, Student Privacy 
• Information Classification 
• Multi-Tenancy Client Protection
• Intellectual Property Non-Disclosure 

After initial setup, your DBA can completely back away.  Developers with basic SQL skills can implement incremental enhancements with ease based on infoSecur’s patterns.  Users need no new skills at all.  Its strength is in its simplicity.

After a few days of initial setup and training, 1-3 analysts can easily manage policies for an entire data environment to be shared with even tens of thousands of diverse users.  It is usually a part-time job.