Automatic Verification of Database-Centric Systems

Software systems centered around a database are pervasive in numerous applications. They are encountered in areas as diverse as electronic commerce, e-government, scientific applications, enterprise information systems, and business process management. Such systems are often very complex and prone to costly bugs, whence the need for verification of critical properties. Classical software verification techniques that can be applied to such systems include model checking and theorem proving. However, both have serious limitations. Indeed, model checking usually requires performing finite-state abstraction on the data, resulting in loss of semantics for both the system and properties being verified. Theorem proving is incomplete, requiring expert user feedback. Recently, an alternative approach to verification of database-centric systems has taken shape, at the confluence of the database and computer-aided verification areas. It aims to identify restricted but sufficiently expressive classes of database-driven applications and properties for which sound and complete verification can be performed in a fully automatic way. This approach leverages another trend in database-driven applications: the emergence of high-level specification tools for database-centered systems, such as interactive web applications and data-driven business processes. We review next a few representative examples. A commercially successful high-level specification tool for web applications is Web Ratio [1], an outgrowth of the earlier academic prototype WebML [20, 17]. Web Ratio allows to specify a Web application using an interactive variant of the E-R model augmented with a workflow formalism. Noninteractive variants of Web page specifications had already been proposed in Strudel [39], Araneus [58] and Weave [40], targeting the automatic generation of Web sites from an underlying database. Highlevel specification tools have also emerged in the area of business process management, concomitantly with an evolution from the traditional process-centric approach towards data awareness. A notable exponent of this class is the business artifact model pioneered in [63, 51], deployed by IBM in professional services. Business artifacts (or simply “artifacts”) model key business-relevant entities, which are updated by a set of services that implement business process tasks. A collection of artifacts and services is called an artifact system. This modeling approach has been successfully deployed in practice [7, 6, 21, 27, 69], and has been adopted in the OMG standard for Case Management [9]. Tools such as the above automatically generate the database-centric application code from the highlevel specification. This not only allows fast prototyping and improves programmer productivity but, as a side effect, provides new opportunities for automatic verification. Indeed, the high-level specification is a natural target for verification, as it addresses the most likely source of errors (the application’s specification, as opposed to the less likely errors in the automatic generator’s implementation). The theoretical and practical results obtained so far concerning the verification of such systems are quite encouraging. They suggest that, unlike arbitrary software systems, significant classes of datadriven systems may be amenable to automatic verification. This relies on a novel marriage of database and model checking techniques, and is relevant to both the database and the computer-aided verification communities. In this article, we describe several models and results on automatic verification of database-driven systems, focusing on temporal properties of their underlying workflows. To streamline the presentation, we focus on verification of business artifacts, and use it as a vehicle to introduce the main concepts and results. We then summarize some of the work pertaining to other applications such as datadriven web services.