Enabling platforms for high - performance computational grids oriented scalable virtual organizations Workpackage 6 Knowledge Services for Intensive Data Analysis and Intelligent Query Answering WP

The activities included in this work package share the common objective of building middleware services for knowledge intensive applications and processes. Specifically, the services we intend to design and implement concern two main topics: • information extraction and knowledge discovery from structured sources (databases or data warehouses) and semi-structured ones (such as web pages or XML documents); • use of extracted information and knowledge within high performance search and query answering high performance services, both for efficiency tuning and for quality of service improvement.. The activities aim at exploiting the lower layers of the Grid architecture in order to implement the functionalities of knowledge discovery and/or to design distributed algorithms (such as data mining algorithms or search engine sub-modules) or to yield modules that can be integrated in the development environment for applications of network and Grid computing. Services that we intend to design are logically structured in three levels over the layer of basic services offered by the toolkits of the grid: Architectural services are developed within the Activity 5, Web Switching. Basic Services are developed within the Activity 1, Basic Services for Knowledge Discovery on Grids, and the activity 3, Resource Discovery and Description. Retrieval services are developed within the activity 4, Retrieval Services. Knowledge Services are developed within the activity 2, Environments and Services for Knowledge Discovery. The services of the KNOWLEDGE GRID are organized into two layers: (1) Core Knowledge-Grid layer and (2) High-level Knowledge-Grid layer. The first layer includes the data management services, by implementing them directly over the basic services offered by the toolkits of the grid. An important data management service is integrating and querying data warehouses distributed on a grid. In particular, methods and tools have been defined to enable interoperability of grid semi-structured information sources using " active " iperlinks in XML documents for service calls to support interaction among sites. In addition, the usage of extensional meta-data has been proposed to provide a synopsis (i.e., an aggregated and approximated description) of the contents of a number of grid information sources, particularly sensors producing a continuous flow of data (data stream). During the first year, we extended and deployed the High level Knowledge Grid architecture, designed a metadata-based information system and a data mining ontology we be used in the system. Knowledge services Retrieval services Basic services Architectural Services 3 The Knowledge Grid is a parallel and distributed software architecture that …