Problems of volumes, time characteristics and physical designing

Distribution of a DB of a class vldb demands more active use of methods of designing of effective physical circuits of the data. It is impossible to build such DB counting on constant reorganization by census in new physical structures. It so for operational DB of a mode oltp, especially it so for terabajtnykh a DB, focused on olap. Ease of procedures of performance of reorganization by the specified method can become "trap" for the designer, is especial - at the first stages of input of a DB in action when its{her} census is still possible{probable} because of incomplete volume.

Expediently at a level of new technologies (application of multivariate structures, indexes of bit displays, etc.) to return to methods of forecasting of operational characteristics of a DB which would allow to plan stability of the physical circuit at least for that time while economic opportunities will not allow to expand external memory of different levels for application of other approaches.

The big growth of volumes of a DB will be accompanied by growth of requirements to their reliability. Control facilities of a DB will directly adjoin to means and methods of designing of a DB owing to constantly carried out process of a redesign of a DB. So, for maintenance steady against refusals of the data possession of control facilities and is necessary for synchronization of geographically carried shadow and reserve databases.

Problem of borders of applicability of two basic methods of designing

In a course of researches and practical designing borders of applicability of two concepts should be determined: designing of a DB as object, is realized separated from applied programs, and object-oriented designing, in which object inkapsuliruet both the data, and methods of their processing.

THE CONCLUSION

Creation of corporate DB in conditions of New System Designing - the activity using many methods of classical designing, but demanding other organization and many additional methods, and also new which would replace some that has been developed 10 and more years ago.

The discipline of designing of a DB in new conditions still is absent. Nevertheless, its{her} beginnings are visible, its{her} elements work in real projects.

According to a principle of preservation of immunity to computer revolutions (see [Zinder95b]) classical methods of designing of a DB should continue to be used, but only in those in areas where they are really useful. The methods of designing considered{examined} in concrete projects corporate IS both a DB, and corresponding tools should be checked on the opportunities to provide function according to requirements of New System Designing.