Figure 1P-22 - Anatomy of a PolicyServer

<font color="#29313b">Of the four subclasses of PolicyApplication, the PolicyServer is the application that operates at the highest level. A PolicyServer is a fundamental building block of a policy-based management system. It represents both a set of core functionality for implementing policy as well as a unit of distribution in a distributed implementation.</font><br/><font color="#29313b">With respect to functionality, a PolicyServer is an entity that is either manufactured or is constructed by integrating the functionality of different PDPs, PXPs, PEPs, and other applications. These other applications provide the logic to manage and control the set of PDPs, PXPs, and PEPs that constitute a PolicyServer. For example, conflict detection and resolution need to be applied. The ContainsPolicyApps aggregation defines the set of PolicyApplications that are contained in this particular PolicyServer. </font><br/><font color="#29313b">The cardinality of this relationship is 0..1 on the PolicyServer side and 3..n on the PolicyApplication side. This means that PolicyApplications can be developed independently of a PolicyServer. However, a PolicyServer must at least have a PolicyDecisionPoint, a PolicyEnforcementPoint, and a PolicyExecutionPoint (PDP, PEP, and PXP, respectively) in order to have the functionality of a PolicyServer. This is indicated by the OCL expression in the figure.</font><br/><font color="#29313b">Therefore, a PolicyServer is in reality an intelligent container that contains its own management logic that is supplemented by the functionality of the PEP, PXP, and PDP applications. It usually corresponds to a single PolicyDomain. While this is not strictly necessary, doing otherwise greatly complicates the resulting management logic.</font><br/><font color="#29313b">Following Figure  shows a simplified view of the primary functions present in a PolicyServer. </font><br/>