Zsolt Kovacs

Meta-data Objects as the Basis for System Evolution

One of the main factors driving object-oriented software development in the Web-age is the need for systems to evolve as user requirements change. A crucial factor in the creation of adaptable systems dealing with changing requirements is the suitability of the underlying technology in allowing the evolution of the system. A reflective system utilizes an open architecture where implicit system aspects are reified to become explicit first-class (metadata) objects. These implicit system aspects are often fundamental structures which are inaccessible and immutable, and their reification as meta-data objects can serve as the basis for changes and extensions to the system, making it self-describing. To address the evolvability issue, this paper proposes a reflective architecture based on two orthogonal abstractions - model abstraction and information abstraction. In this architecture the modeling abstractions allow for the separation of the description meta-data from the system aspects they represent so that they can be managed and versioned independently, asynchronously and explicitly. A practical example of this philosophy, the CRISTAL project, is used to demonstrate the use of meta-data objects to handle system evolution.

Pattern reification as the basis for description-driven systems

One of the main factors driving object-oriented software development for information systems is the requirement for systems to be tolerant to change. To address this issue in designing systems, this paper proposes a pattern-based, object-oriented, description-driven system (DDS) architecture as an extension to the standard UML four-layer meta-model. A DDS architecture is proposed in which aspects of both static and dynamic systems behavior can be captured via descriptive models and meta-models. The proposed architecture embodies four main elements – firstly, the adoption of a multi-layered meta-modeling architecture and reflective meta-level architecture, secondly the identification of four data modeling relationships that can be made explicit such that they can be modified dynamically, thirdly the identification of five design patterns which have emerged from practice and have proved essential in providing reusable building blocks for data management, and fourthly the encoding of the structural properties of the five design patterns by means of one fundamental pattern, the Graph pattern. A practical example of this philosophy, the CRISTAL project, is used to demonstrate the use of description-driven data objects to handle system evolution.

Handling Evolving Data Through the Use of a Description Driven Systems Architecture

Traditionally product data and their evolving definitions, have been handled separately from process data and their evolving definitions. There is little or no overlap between these two views of systems even though product and process data are inextricably linked over the complete software lifecycle from design to production. The integration of product and process models in an unified data model provides the means by which data could be shared across an enterprise throughout the lifecycle, even while that data continues to evolve. In integrating these domains, an object oriented approach to data modelling has been adopted by the CRISTAL (Cooperating Repositories and an Information System for Tracking Assembly Lifecycles) project. The model that has been developed is description-driven in nature in that it captures multiple layers of product and process definitions and it provides object persistence, flexibility, reusability, schema evolution and versioning of data elements. This paper describes the model that has been developed in CRISTAL and how descriptive meta-objects in that model have their persistence handled. It concludes that adopting a description-driven approach to modelling, aligned with a use of suitable object persistence, can lead to an integration of product and process models which is sufficiently flexible to cope with evolving data definitions.

The Design of an Engineering Data Warehouse Based on Meta-Object Structures

Large scale engineering and scientific projects demand product and workflow management which may require integration and/or distribution over many separate organisations. The integration of such ‘islands of information’, which ultimately forms the basis of so-called ‘virtual enterprises’, is heavily dependent on the flexibility and accessibility of the data model describing the enterprise’s repository. The model must provide interoperability and reusability so that a range of applications can access the enterprise data. Making the repository self-describing ensures that knowledge about the repository structure is available for applications to interrogate and to navigate around for the extraction of application-specific data. Herein a large application is described which uses a meta-object based repository to capture product and workflow data in an engineering data warehouse. It is shown that adopting a meta-object approach to repository design provides support for interoperability and a suitable environment on which to build data mining applications.

Designing Traceability into Big Data Systems

Providing an appropriate level of accessibility and traceability to data or process elements (so-called Items) in large volumes of data, often Cloud-resident, is an essential requirement in the Big Data era. Enterprise-wide data systems need to be designed from the outset to support usage of such Items across the spectrum of business use rather than from any specific application view. The design philosophy advocated in this paper is to drive the design process using a so-called description-driven approach which enriches models with meta-data and description and focuses the design process on Item re-use, thereby promoting traceability. Details are given of the description-driven design of big data systems at CERN, in health informatics and in business process management. Evidence is presented that the approach leads to design simplicity and consequent ease of management thanks to loose typing and the adoption of a unified approach to Item management and usage.

Promoting Reuse through the Capture of System Description

One of the main drivers in object-oriented design for information systems is the need for the reuse of design artifacts in handling systems evolution. To be able to cope with change, systems must have the capability of reuse and to adapt as and when necessary to changes in requirements. To address the issues of reuse in designing evolvable systems, this paper proposes a so-called description-driven system architecture. The proposed architecture is based on a twodimensional design approach founded on the adoption of a multilayered modeling architecture and on a reflective meta-level architecture. This paper discusses the need for capturing holistic system description when modeling large-scale distributed systems and the role of reflection as a method to cater for reuse in systems evolution. A practical example of the application of this design philosophy, the CRISTAL project, is used to demonstrate the reuse of descriptiondriven data objects to provide for evolution.

Analysis traceability and provenance for HEP

This paper presents the use of the CRISTAL software in the N4U project. CRISTAL was used to create a set of provenance aware analysis tools for the Neuroscience domain. This paper advocates that the approach taken in N4U to build the analysis suite is sufficiently generic to be able to be applied to the HEP domain. A mapping to the PROV model for provenance interoperability is also presented and how this can be applied to the HEP domain for the interoperability of HEP analyses.

Reifying design patterns to facilitate systems evolution

In the Web age systems must be increasingly flexible, reconfigurable and adaptable in addition to being developed rapidly. As a consequence, designing systems to cater for change is becoming critical to their success. Allowing systems to be self-describing or description-driven is one way to enable this. To address the issue of evolvability in information systems, this paper proposes a pattern-based description-driven architecture. The proposed architecture embodies four pillars – firstly, the adoption of a multi-layered and reflective meta-level architecture, secondly, the identification of four modeling relationships that must be made explicit to be examined and modified dynamically, thirdly the identification of five patterns which have emerged from practice and have proved essential in providing reusable building blocks, and finally the encoding of the structural properties of these design patterns by means of one pattern, the Graph pattern. A practical example of this is cited to demonstrate the use of description-driven data objects in handling system evolution.

A component based approach to scientific workflow management

CRISTAL is a distributed scientific workflow system used in the manufacturing and production phases of HEP experiment construction at CERN. The CRISTAL project has studied the use of a description driven approach, using meta-modelling techniques, to manage the evolving needs of a large physics community. Interest from such diverse communities as bio-informatics and manufacturing has motivated the CRISTAL team to re-engineer the system to customize functionality according to end user requirements but maximize software reuse in the process. The next generation CRISTAL vision is to build a generic component architecture from which a complete software product line can be generated according to the particular needs of the target enterprise. This paper discusses the issues of adopting a component product line based approach and our experiences of software reuse.

Explicit modeling of the semantics of large multi-layered object-oriented databases

Description-driven systems based on meta-objects are an increasingly popular way to handle complexity in large-scale object-oriented database applications. Such systems facilitate the management of large amounts of data and provide a means to avoid database schema evolution in many settings. Unfortunately, the description-driven approach leads to a loss of simplicity of the schema, and additional software behaviour is required for the management of dependencies, description relationships, and other Design Patterns that recur across the schema. This leads to redundant implementations of software that cannot be handled by using a framework-based approach. This paper presents an approach to address this problem which is based on the concept of an ontology of Design Patterns. Such an ontology allows the convenient separation of the structure and the semantics of database schemata. Through that, reusable software can be produced which separates application behaviour from the database schema.