[Paper Review] Design Patterns for Description-Driven Systems

Goff, J. -M. Le, Ian Willers+2 more|ArXiv.org|Jul 21, 1999

Business Process Modeling and Analysis16 references3 citations

TL;DR

This paper proposes a description-driven system that unifies Product Data Management (PDM) and Workflow Management (WfM) through a common meta-model, leveraging existing design patterns and identifying gaps for integration. It demonstrates how multi-layer architectures and pattern enrichment enable lifecycle information sharing from design to production.

ABSTRACT

In data modelling, product information has most often been handled separately from process information. The integration of product and process models in a unified data model could provide the means by which information could be shared across an enterprise throughout the system lifecycle from design through to production. Recently attempts have been made to integrate these two separate views of systems through identifying common data models. This paper relates description-driven systems to multi-layer architectures and reveals where existing design patterns facilitate the integration of product and process models and where patterns are missing or where existing patterns require enrichment for this integration. It reports on the construction of a so-called description-driven system which integrates Product Data Management (PDM) and Workflow Management (WfM) data models through a common meta-model.

Motivation & Objective

• Address the longstanding separation of product and process information in data modeling.
• Enable seamless information sharing across the enterprise lifecycle—from design to production.
• Identify where existing design patterns support or fall short in integrating PDM and WfM models.
• Develop a description-driven system architecture based on a shared meta-model for PDM and WfM.
• Provide a framework for extending or enriching design patterns to support integrated system modeling.

Proposed method

• Adopt a multi-layered architectural approach to decouple description logic from implementation.
• Define a common meta-model that serves as a shared foundation for both PDM and WfM data models.
• Map existing design patterns to the integration of product and process models in the meta-model.
• Identify missing or insufficient patterns for handling complex interdependencies between product and process data.
• Propose enhancements to existing patterns to support dynamic, lifecycle-aware system descriptions.
• Validate the approach through a case study at CERN, demonstrating integration in a real-world enterprise context.

Experimental results

Research questions

• RQ1 How can product and process data models be unified within a single, coherent meta-model?
• RQ2 Which existing design patterns are applicable to the integration of PDM and WfM systems?
• RQ3 What gaps exist in current design patterns when applied to description-driven systems with lifecycle requirements?
• RQ4 How can the meta-model support dynamic evolution of system descriptions across design, production, and maintenance phases?
• RQ5 What architectural patterns enable scalable and maintainable integration of PDM and WfM in enterprise systems?

Key findings

• A common meta-model successfully unifies PDM and WfM data models, enabling consistent representation across the system lifecycle.
• Existing design patterns such as Strategy, Template Method, and Composite were found applicable but required extension for full integration.
• The study identified specific gaps in pattern support for lifecycle-aware data dependencies between product and process models.
• Enriched patterns enabled dynamic configuration and runtime adaptation of system descriptions, improving maintainability.
• The description-driven system prototype demonstrated effective data sharing across design and production phases in a CERN-based context.
• The approach reduced data redundancy and improved consistency by enforcing a single source of truth through the meta-model.