A Flexible Grid Infrastructure for Data Analysis
In this paper we present a novel infrastructure for data analysis based on the mechanism of updatable views. The views are used in two ways: as wrappers of local servers, which allow one to adopt local schemata to warehousing requirements, and as a facility for data integration and transformation to a federated canonical data model. Such updatable views present virtual data to the clients, but they themselves are also objects that can have methods and other features. The advantage of our approach is a simple architecture and a user-friendly query language for definition of views and for writing applications. Thus our approach can compete with Web Services, CORBA and other middleware technologies in terms of universality, flexibility and maintainability. This approach is based on the Stack-Based Approach to query languages, in which the query language is treated as a kind of programming languages.
Unable to display preview. Download preview PDF.
- 1.Oracle 10g: Infrastructure for Grid Computing, Oracle White Paper (September 2003)Google Scholar
- 2.OMG: OMG CORBATM/IIOPTM Specifications (2002), http://www.omg.org/technology/documents/corbaspeccatalog.htm
- 6.Subieta, K., Kambayashi, Y., Leszczyłowski, J.: Procedures in Object-Oriented Query Languages. In: Proc. of 21-st VLDB Conf., pp. 182–193 (1995)Google Scholar