Content Source Description and Selection Services (CSDS)
Overview
The Content Source Description (CSD) supports the execution of content-based queries against a number of target content sources, such as collections associated with DILIGENT indices ( internal collections) and collections which interface DILIGENT via dedicated search front ends ( external collections). The service does that by providing estimations of "the goodness" of the content source based on information such as partial content indices, summary content indices, or result traces for training or past queries. The generation and maintenance of such descriptions is the prime responsibility of the CSD service.
The Content Source Selection (CSS) is a DILIGENT service which supports the execution of content-based queries against a number of target content sources, such as collections associated with DILIGENT indices (internal collections) and collections which interface DILIGENT via dedicated search front ends ( external collections). The CSS service limits the routing of queries to those sources which appear to be the best targets to their execution. "Goodness" here includes the relevance of the content, the sophistication of retrieval engines, and/or the costs of the monetary costs associated with query execution.
Features
- Internally homogeneous and robust, fully developed in Java
- Inherently Service Oriented
- Provide comprehensive server-sided logging and exception handling
- Reuse several underlying DILIGENT platform logical layers and external technologies thus avoids duplication of effort and promotes stability.
- Handle efficiently the plurality and dynamicity of resources available on a DILIGENT grid infrastructure.
- Are modular, open and extendible, easily extendible to capture application domain specific needs
- Exploits indexing facilities provided by external services and supports different indexing strategies
- Provides persistency of content source descriptions with content management system
- Provides different access modes to description information based on client requirements
- Provide extensive support for different selection strategies
- Provides a selection strategy based on the CORI algorithm
Read More
Related Components & References
The CORI Algorithm (CORI Paper: James P. Callan, Zhihong Lu, and W. Bruce Croft. Searching distributed collections with inference networks. In SIGIR '95: Proceedings of the 18th annual international ACM SIGIR conference on Research and development in information retrieval, pages 21–28, New York, NY, USA, 1995. ACM Press.)