JoWUA

Abstract

Process-Aware Information Systems (PAIS) enable the definition, execution, and management of business processes. Typically, processes are specified by control flow, data flow, and users or services, authorized to execute process tasks. During process execution, it is often necessary to access sensitive data such as patient or customer information. To secure this confidential data, the use of security policies becomes an essential factor for the application of PAIS in practice. In general, PAIS security policies are specified based on access rules and authorization constraints. On top of these rules, context policies referring to data, location, or time might pose restrictions. Over the years, several approaches for modeling and enforcing security policies in PAIS have appeared. Many of them restrict security policy specification to access rules and authorization constraints, but neglect additional properties such as context information. As a further limitation, security policies are often defined in a heterogeneous way: whereas access rules are mostly defined at process task level leading to a merge of process logic and security aspects, additional policies such as authorization constraints are defined separately from the process logic. Consequently, security policies are not stored and managed centrally, but are rather distributed over different PAIS components, for example, the process model repository or the organizational model manager. In this paper, we introduce the formal concepts behind our SPRINT approach that aims at the consequent separation of security policies and process logic. Specifically, the SPRINT security policy data model and design methodology based on the concepts of responsibilities, permissions, and constraints will be provided. The concepts are evaluated based on a comparison with existing PAIS and a demonstration of the SPRINT prototype. The goal is to unify diverse security policies in different PAIS subsystems, to make security policies independent of these subsystems in order to restrain complexity from process modeling and evolution, and to allow for comprehensive security policy development and maintenance.