Skip to content
Software Architecture Guide

Data Architecture Principle

Explanation

Section titled “Explanation”

A data architecture principle is a basic rule for handling data. It guides design choices about data. The examples below follow TOGAF-style data principles (see the article “TOGAF”).

Typical principles

  • Data is an asset. Data is valuable. It supports decisions and helps the organization operate effectively.
  • Data is shared. Data is shared across enterprise functions and organizations. The rationale: efficient sharing can improve decisions, because people rely on a single accurate source of information.
  • Data is accessible. Users can reach the data they need to do their jobs.
  • Data has a trustee. Each data element has a trustee who is accountable for its quality. One implication is quality control measures that keep data accurate and current.
  • Data is secure. Data is protected from unauthorized use and disclosure.

Implications and governance

A principle also lists what follows from it. One implication ties to governance: functional data administration responsibilities should be clearly assigned. Without a named owner, a principle is only a wish.

Principles in arc42

arc42 is a documentation template. It treats principles as one kind of cross-cutting concept. Other kinds are approaches, rules, tactics, and strategies (see “Crosscutting Concept”). arc42 says concepts form the basis for conceptual integrity, meaning the architecture stays consistent and homogeneous. It also advises documenting at least the business or domain data model as a concept.

For principles in general, see “Architecture Principle”.

Example

Section titled “Example”

A retail company writes down three data principles. First: customer data is shared across sales and support. Second: each data element, such as the customer address, has a named trustee who answers for its quality. Third: customer data is protected from unauthorized use and disclosure. Each principle gets a short statement, a reason, and a list of consequences. One consequence is that someone must be assigned to administer the customer data.

Trade-offs

Section titled “Trade-offs”

The facts here do not describe costs in detail. The principles do pull in different directions. Sharing and accessibility make data easier to reach. Security limits who may use it. Principles also create work: quality control measures must be put in place, and administration duties must be assigned.

Checks

Section titled “Checks”
  • Each important data element has a named trustee accountable for its quality.
  • Quality control measures exist to keep data accurate and current.
  • Data administration responsibilities are clearly assigned.
  • A principle states that data is protected from unauthorized use and disclosure.
  • The business or domain data model is documented as a concept.
Section titled “Related”

Sources

Section titled “Sources”