Information Server Architecture

Verity's search technology is embedded in all of its search products. The basic components of this search technology, as it is implemented in Information Server, are described in this section.

Information Server Architecture

Information Server shares a common architecture with a number of other Verity products that allows seamless communication with your Web server. The Information Server architecture can be portrayed as follows:

The Information Server architecture allows the component framework to communicate seamlessly with your Web server regardless of the communication protocol it uses. The following protocols are supported:

NSAPI for Netscape servers
ISAPI for Microsoft Internet Information Server
CGI for other servers, or if you prefer to use it for Netscape or Microsoft

The following are some of the Information Server components that use this architecture to communicate with your Web Server:

Index - The index component indexes documents in numerous formats, storing the index information in units called collections. A Verity collection consists of indexed data about your documents, not the actual documents. The system administrator uses the Verity spider to submit indexing tasks. The Verity spider is available in GUI and command-line forms.
Search - The search component parses the query the user enters and sends it to the search engine. The search component can be configured to use the optional query filter component, described below.
Query Filter (optional)- The query filter allows you to pre-filter queries. For example, you may wish to restrict the query to a specific collection, or filter out certain documents for security purposes.
Results List- The results list component formats the results list that is returned as a result of a search.
Document View - The document viewing component allows you to customize the display of the document, including highlight navigation, which is presented when the user selects a retrieved document from the results list.

Verity Spider

The Information Server indexing component, or spider, can create collections from a variety of document types, including:

ASCII text documents. This includes text documents and documents in internet e-mail or usenet news format.
HTML documents.
Adobe Acrobat PDF documents.
WYSIWYG documents. This includes documents supported by the KeyView Filter Kit, including Microsoft Word, WordPerfect, and most popular word processing, spreadsheet, and presentation graphics formats.

The Information Server indexing component can index all of these various document types into a single meta-collection. While internally each document type is indexed into a separate collection, the end result appears to be a single collection. This greatly simplifies collection management and making them available to users.

Collections

A collection represents a group of related documents that can be searched. A collection includes a number of elements that together facilitate rapid searching and retrieval of documents:

Indexes optimized for searching
References to actual document locations
Field values that may reference such standard document elements as title or date

A collection does not contain the actual documents to be searched, but only a reference to the location of the documents.

Copyright © 1998, Verity, Inc. All rights reserved.