Understanding policy models

There are several policy models created and maintained by different roles, used for different purposes, each with different lifecycle concerns.

How you author, test, execute, and maintain a growing set of policies in the Fabric is important in the course of your project development.

A Fabric Policy can be decomposed into Policy Conditions and Policy Assertions. A Policy expresses "if conditions, then assertions." The conditions of a policy include an effective date range, a Policy Target, and expressions that involve either concepts (context conditions) in the Business Services Repository, or content-based assertions populated from the context.

Policy Assertions are instances of reusable concepts that have been modeled for a specific domain. A policy-based decision uses the set of assertions from a composite policy as inputs to decision making. A composite policy is derived by merging the assertions of all applicable policies. The set of applicable policies is determined by finding all policies whose policy conditions are satisfied given the information in the context.

A Natural Ordering System orders determines which assertion to use when an assertion of a given type is expressed by multiple policies. The policy target is ranked by its natural order within a natural ordering system. Typically a system orders domain concepts from most general to most specific.

For example, Organization is considered more general than User. Assertions are chosen from the policy with the most specific natural order unless the assertion is explicitly locked at higher natural order. When assertions are at the same natural order, the priority of a policy can be used to pick the winner. Finally, if no winner can be chosen, this is called a Policy Conflict.

Related concepts
Establishing service levels
Working with a dynamic assembly simulation
Related tasks
Creating a policy
Related reference
Working with the policy editor