|Title |Introduction |Integrating App's |Integrity Maintenance |AOODB Model |Backward Propagation |HVC |Constraint Maintenance |Conclusion |References|
8 Conclusion
In existing CIM environments, verification and enforcement of engineering constraints are performed by heterogeneous engineering applications. The application-oriented integrity maintenance scatters the constraints in the applications and makes it difficult to manage and exchange them consistently. The CIM systems based on this integrity maintenance approach have suffered from the engineering semantics loss that makes translated data inconsistent when it is exchanged between the applications, hence causing tedious and time-consuming reworks.
As a solution of the engineering semantics loss, a framework that employs the centralized integrity maintenance has been proposed. The framework enables applications to share data and its constraints efficiently, since they are managed together. The framework, therefore, can help realize the information driven-CIM that integrates heterogeneous applications based on the information exchanged. For the framework to achieve the goal, a database that can manage both the data and its constraints should be employed.
The framework includes an active object-oriented database that supports user-defined integrity constraints. The rules for constraint enforcement, implemented with the trigger constructors, are stored in the database so that the applications may share the object, methods and rules consistently. Further, triggers with methods allow to specify complex procedural constraints on engineering objects. The object-oriented data model also enables to express complex entities in engineering domain.
Existing active object-oriented databases, however, cannot properly support certain constraints that are specified on a set of classes that form a class composition hierarchy. Those are the cases where the constraints must be maintained in the forward direction along the class composition hierarchy as well as in the backward direction. We have named these kinds of problems as "backward propagation problems" and investigated several alternatives to resolve them using currently available techniques. Based on one of the approaches which uses virtual classes, a new constructor, called CONSTRAINTCCH has been proposed to support integrity maintenance in a class composition hierarchy. It uses only resources from an active object-oriented database so that the applications may be integrated without additional facilities for constraint maintenance. Moreover, the virtual class in the backward propagation approach, which we call Horizontal Virtual Class, enables efficient representation and management of constraints, using participating objects' identifiers and methods for constraints.
A prototype design system based on the framework and the constructor for the backward propagation has been implemented to support control panel design. It demonstrates that the framework is able to realize an integrated environment where all the applications can share engineering data as well as its constraints without additional facilities for the constraint maintenance.
Our research can be compared with other approaches in two aspects. First, in terms of the integration of applications, the proposed framework provides an integrated environment where applications can share data, methods, and rules for integrity constraints. Furthermore, since the database automatically detects and repairs the change of objects that may violate specified integrity constraints, the applications to be integrated are free from the detailed execution logic for constraint enforcement. This will enable heterogeneous applications to be integrated in an information driven CIM environment without concern about the constraints. Secondly, in terms of the constraint maintenance in active object-oriented databases, the backward propagation approach is able to specify constraints on multiple classes in a class composition hierarchy. Since it uses virtual classes for the propagation of the constraints, the approach efficiently propagates the constraints without duplicated data for the constraint maintenance. In addition, there is, known to date, no report on the investigation about the backward propagation in active object-oriented databases with the systematic resolution capability such as CONSTRAINTCCH.
One of the future research direction is to extend the proposed framework to be based on multi databases. In today's corporate CIM environment, managing engineering information in large and isolated databases designed for specific platforms is not reasonable. Many engineering systems including CIM systems are being constructed on a set of heterogeneous distributed databases. The need for a transparent constraints on the databases has also been recognized. These factors require the framework employ a multi database architecture that provides a unified constraint on heterogeneous distributed databases.
We have provided the EXPRESS specification language and its translator as a component of the framework in Section 7.1. The current implementation only translates the local rules of EXPRESS that concerns constraints within an entity. The global rule for constraints on multiple entities is not implemented due to the complexity from the flexible expressions of the rules. Since the backward propagation is a kind of the global rule, if the automatic detection of the constraint for the backward propagation is prepared in the framework, the constructor for the backward propagation can be integrated into the global rule.
For the backward propagation, we leave the case where the participating classes form a class inheritance hierarchy as a further work. To support the backward propagation along the inheritance hierarchy, the problems typically associated with the class inheritance, such as change propagation to subclasses, should be considered. The research of the propagation of constraints on class inheritance hierarchy associated with the existing work on class composition hierarchy will provide a comprehensive investigation about the constraint propagation on the multiple classes in active object-oriented databases.
Korean Engineering Databases ¨Ï copyright Namchul Do, 1996