Resumen de MARBLE: modernization approach for recovering business processes form legacy informaticon systems
Ricardo Pérez Castillo
This PhD Thesis proposes and validates MARBLE, a framework and a set of framed techniques with which to recover and rebuild the business process models that are actually supported by legacy information systems. MARBLE, which can be considered as a business process archeology method, makes it possible to preserve the business knowledge buried in legacy information systems, thus allowing it to be effectively used during software modernization. MARBLE is a generic framework based on ADM (Architecture-Driven Modernization), which uses the KDM (Knowledge Discovery Metamodel) to represent the information retrieved in a common and standard manner. MARBLE proposes a set of models at four different abstraction levels according to the model-driven development principles, along with three model transformations between those levels to progressively reduce the gap between legacy information systems and business process models. This thesis proposes two particular techniques framed in MARBLE. The first technique statically analyses the source code and represents the embedded knowledge in a KDM repository, which is then analyzed by applying a pattern matching technique to discover business process models, while the second technique analyzes the source code at runtime to obtain event logs which are then analyzed to retrieve business processes. The research methodology includes the action research method combined with research based on case studies. The proposed framework has been validated with several case studies involving real-life industrial legacy information systems in different domains (health-care, education, e-administration, and so forth).
