Direct Indexing vs. Federated Search

Search Server 2008, the successor to SharePoint Server 2007 for Search, will support a new type of search architecture called federated search.

Like its predecessor, Search Server 2008 will support direct indexing (upper left). In this model, Search Server’s indexing service crawls data sources, such as SharePoint sites, file servers, Exchange Public Folders, and Lotus Notes databases, using software components called iFilters to read various types of files (e.g., Office documents, Adobe PDF files) and components called protocol handlers to discover and transfer data over various types of data connections (e.g., MAPI). It populates its index with data crawled from these sources, and users conduct searches against this index. Direct indexing requires Microsoft, ISVs, or customers to create the iFilters and protocol handlers for the data sources being indexed. SharePoint Server 2007 supports many data sources out of the box, but most corporate business applications, such as ERP applications, were not supported, requiring considerable extra effort on the part of customers.