Federating Neuroscience Databases [chapter]

Wen-Hsiang Kevin Liao, Dennis McLeod
2001 Computing the Brain  
There are three key aspects of sharing and interconnection in a federated database environment: information discovery, semantic heterogeneity resolution, and system-level interconnection. Although the focus here is on the system-level interconnection process of the three key aspects described above, we also provide brief discussions on the other two aspects and summarize our approaches to those two issues. Our approach to support information sharing among databases is based on the import/export
more » ... paradigm. In this paradigm, a component database of a federation decides the portion of its database to be exported and offers the methods to others on how the information can be shared. Other component databases that are interested in using the information can import the information using one of the sharing methods provided by that component database. Information sharing is thus reached based upon the agreements (contracts) between each pair of component databases in which sharing is needed. A set o f sharing primitives/tools is designed to support these sharing activities in a uniform manner. The sharing primitives/tools let users to focus on how their information could be shared efficiently rather than the underlying implementation. This enables users of a component database to utilize not only the Liao and McLeod: Federating Neuroscience Databases January 2000 2 information in their own component database but also the information in other component databases of the same federation. component database 2 Liao and McLeod: Federating Neuroscience Databases January 2000 3 There are three key aspects of sharing and interconnection in a federated database environment. These may be viewed in the context of a given component database (C), which intends to import information from and/or export information to other component databases. Information discovery: The information discovery process pertains to finding out what information can be shared in the first place. From the viewpoint of a component database C, the main concern is to discover and identify the location and content of relevant, non-local information units in the federation. Semantic heterogeneity resolution: Various kinds of diversity may exist in the autonomous component databases in a federation. The similarities and differences between component C's information and relevant non-local information need to be resolved so that they can be integrated and shared. System-level interconnection: Based upon existing networking technologies, mechanisms and its implementation must support actual sharing and transmission of information to and from C and other components in the federation. The focus of this chapter is on the system-level interconnection process of the three key aspects described above. In addition, we also provide brief discussions on the other two aspects and summarize our approaches to those two issues. The remainder of the chapter is organized as follows. Section 2 briefly describes the dynamic classificational ontology approach to information discovery problem. Section 3 outlines the spectrum of semantic heterogeneity and introduces the approach of meta-data implantation and stepwise evolution. Section 4 discusses the issues related to system-level interconnection. Section 5 describes the characteristics of sharing patterns. Section 6 layouts the system architecture of our approach in providing sharing primitives/tools. We conclude the chapter with a sharing scenario using neuroscience databases in section 7. Liao and McLeod: Federating Neuroscience Databases January 2000 4
doi:10.1016/b978-012059781-9/50016-x fatcat:74jhgnez4jedvn342zrjqmdome