(A) change management and view synchronization framework in distributed model management systems

Abstract

Large-scale model management systems are increasingly in demand to support groups of users in a collaborative work environment. Users of such systems usually share common models, but approach them with different user-views to meet individual needs and skills. Moreover, in distributed model management environments, there exist dependent models stored in modelbases regardless of their locations. In this capacity, since models change as does the reality that is captured in them, so should their dependent models and user-views. This calls for a mechanism to manage such changes and synchronize the dependent models and user-views so that the integrity of the modeling environment can be maintained. In this dissertation, we propose a change management and view synchronization framework in distributed model management systems, which (1) manages the dependency relationships not only between the base models and their dependent models but also between the shared models and their dependent user-views, and (2) propagates the changes occurred in the model to its dependents immediately and automatically. To do so, we identify the change management requirements in distributed modeling environments and specify the effects of such changes to the dependents. Since the dependent models stored in a modelbase are persistent objects concurrently accessed by multiple users, to guarantee the change reflection task to them securely, the two-phase commit protocol is adopted in the framework. Specifically, we define the concrete object-oriented constructs and processes to internalize the proposed change management mechanisms in a distributed model management system, and utilize the petri-net theory to formalize and demonstrate the proposed framework. In the framework, the generic model concepts are adopted for structuring a modelbase with object-oriented constructs, and the object-oriented database management systems (ODBMS) for combining the constructs of model management and those of cha...