Improving water information management through information modelling

The Bureau of Meteorology has responsibility for collecting, holding, managing, interpreting and disseminating Australia’s water information. Given this mandate, the Bureau is developing the Australian Water Resources Information System (AWRIS) to receive and manage water data and related information, and to support the production and dissemination of a variety of water information ‘products’ and services. Traditionally, as business needs for data and information changed and developed over time, it would be necessary to re-design data models, databases, analysis tools and products to ensure continuity in data and information supply. An ‘information modelling’ approach that aims to address long term maintenance and evolution is being trialled as part of development of AWRIS. This approach uses Model Driven Architecture/ Model Driven Development (MDA/MDD) within an enterprise architecture framework to maximise reusability, scalability, platform independence, speed-of-delivery, change control and governance. The approach entails a philosophy and method of development, and a tool (Sparx Enterprise Architect) for translating formal platform-independent models to platform dependent scripts, schemas or code. In its most ideal form, MDA/MDD offers: • A formal and constrained approach to specification of business needs, in the form of models written in UML • Integrated documentation as the model is loaded and described using Enterprise Architect • Rapid and agile development, with automated transformation of models into operational schemas, services and/or code • Increased collaboration between domain and technical experts, through clearer sharing of needs, more common language for communication, faster turnaround on changes, and better feedback In practice there are a number of challenges to adoption this approach, both in concept, governance and operation. Conceptual challenges have been tackled through use of an architectural framework the Zachman Framework – that provides a matrix structure for relating project drivers, concepts and ‘artefacts’ (e.g. documents, code). From top to bottom the framework passes from Scope (Context) at the highest level through Business (Concept), System (Logic), Technology (Physics), Component (Assemblies), to Operational (Instantiation) level. For AWRIS the scope might be defined by the requirements of the Water Act 2007, while the operational level might be a data delivery service. Across each level within the framework (e.g. across the ‘Technology’ level) lie the fundamental communication elements, covering the ‘What’ (Inventory), ‘How’ (Process), ‘Where’ (Network), ‘Who’ (Organisation), ‘When’ (Timing) and ‘Why’ (Motivation). Governance and operation of the MDD/MDA approach have provided challenges that include variations in application due to unfamiliarity with UML and the Zachman framework, model repository governance and change control as use and user numbers expand, and inconsistencies between predefined schema, model and documentation, confounding the effectiveness of the transformation from platform independent to platform dependent artefacts.