Chapter 13. Search Engine

Searching has become a serious challenge for the enterprise. Your portal by now contains a plethora of content, ranging from structured data in relational databases to unstructured data such as web content and documents. The data is stored in a number of formats and scattered across multiple locations. In most cases, the metadata is incomplete and inaccurate. For instance, how many users do you know who accurately fill in the title and author in the properties of Word documents? I confess that I often end up copying documents and moving sections around so much that I am surprised to see the contents of the document properties. With the increase in team collaboration online, critical documents may be located in collaboration sites, on file shares, and in email. In some enterprise architectures, each format requires a different search technique, making unified searches that cut across data sources impractical . A key challenge is to make this data accessible to all the categories of users of the portal.

This chapter shows how to plan, install, configure, and manage search capabilities for the portal. Most of these search technologies are full-text searches for unstructured data rather than structured queries in a relational database. Searches for both structured and unstructured data sources are needed for most portals.

Microsoft has developed full-text search technology that is implemented in several server products. It has made a significant investment in research and development for search (see Microsoft Researches Search Technology). For instance, full-text search capability is included in Indexing Service, SQL Server, Exchange Server, Office, and SharePoint Portal Server. The features of the individual products vary, but the core functionality is quite similar, and many of the search concepts carry over from one product to another. Indeed, the behavior of searches can be expected to converge as Microsoft makes better use of its research and development investment in search technology.

This chapter explains how full-text search can improve your portal and covers important concepts for search, a broad overview of the search engine architecture, and a comparison of the Microsoft server products that provide this feature. The primary emphasis in this chapter is on the search engine included in Microsoft SharePoint Portal Server because it is the most advanced search product that Microsoft offers. Windows SharePoint Services contains its own search feature as well, but it is not as powerful or flexible as SharePoint Portal Server. SharePoint Portal Server includes several features to make searches faster and more productive, including:

• Ability to search multiple information stores through a single search engine

• Search for not only documents but people (profiles), sites, and topics

• Scheduled index updates

• Full-text searches for documents and web pages

• Metadata searches for document properties

• Topic categories to browse search results

• Automatic categorization of documents

• Best Bet classification for highly relevant search results

• Multiple options for grouping and presenting search results by site, topic, author, and date

• Simple and advanced search

• Scalable indexing and searching architecture

Other SharePoint features that are discussed in other chapters complement search capabilities. For instance, alerts help portal users stay up to date on new search results, and taxonomies are important tools for searching and browsing (see Chapter 10). The white paper "Microsoft Full Text Search Technologies" (June 2001) provides an excellent overview of Microsoft search technologies. You can find it at www.microsoft.com/technet/prodtechnol/sppt/sharepoint/evaluate/featfunc/ mssearch .asp.