We have described an experimental implementation of Diorama, a cooperative information system based on the DIOM approach [19] to support interoperability in internet environments with heterogeneous and autonomous components. Diorama is constructed using web interfaces, Oracle, and CGI scripts. A demo is accessible through Web browsers on the Internet (see Section 4.6).
The main contribution of this paper is a concrete demonstration of the practical feasibility of implementing DIOM. Being a powerful model, DIOM supports the USECA properties (Uniform access, Scalabiltiy, Evolution, Composability, and Autonomy) of a heterogeneous information system. DIOM facilitates the interconnection between a variety of information consumers and similarly variable information producers. Services that Diorama provides include producer information source registration, metadata library management, query routing, query decomposition, subquery transformation and execution, and result assembly. Highlights of the current implementation include:
We have made significant progress towards the second phase of the Diorama implementation, which will focus on three main aspects: (1) the creation of consumer's profile, the metadata catalog management, described in [17], and the producers' profile organization; (2) the realization of the query decomposition algorithm, described in [21] and query assembly operations based on consumer's profile, producers' registration data, and domain-specific ontology or thesaurus; and (3) the algorithms for generating optimized parallel access plans. We are also interested in developing generic wrapper functions that can be used to generate wrapper instances for various types of data repositories.