Microsoft SharePoint Server 2010 builds on the existing capabilities of SharePoint Server to deliver a more robust enterprise content management and collaboration solution. One feature of SharePoint Server that has seen a radical change is SharePoint Search. Microsoft has introduced several changes in the search architecture to connect people with information more easily than before. You can use SharePoint 2010 Search to conveniently search content and people, and even develop various search-driven applications.
The following list describes major changes in the Search architecture introduced in SharePoint 2010:
This server is used to host the server as in SharePoint Server 2007. However, the index is divided and a new concept has been introduced called the index partition. This was important as there was a single SSP (Shared Service Provider) associated with one index in the previous version. The SSP was a single point of failure.
This server has the same function, i.e. copying content from the web server and creating index. It brings to the table a new thing called 'Crawl Components'.
This server hosts crawlers and each crawler has an associated crawl database. These crawlers help you crawl content based on the information in the crawl database. The server also has a search administration content. You should have at least one server in the farm to perform the index role. Two or more index servers usually allow for improving the crawl performance and keeping up with the capacity requirements.
Index partitions provide the ability to create an index using multiple content sources that are scattered across various query servers. Further, the concept of index partitions allow for fault tolerance. This is achieved by placing index partitions on multiple query servers such that you do not lose the entire index if one server stops functioning. This poses the need to have at least two query servers to achieve some degree of fault tolerance.
Search Admin Component
The primary role of Search admin component is to implement changes in topology and provide easy administration for the search instance. You cannot have more than one Search administration component for a search service application.
This server can support both mirrored and clustered configuration. It hosts all search related databases in your farm. In order to scale out for capacity requirements, you can have more servers and disks. The database server primarily hosts three databases, i.e. property database, crawl database, and search administration database. The property database contains the properties of the data crawled through crawlers and the crawl database.
SharePoint 2010 Search features many enhancements that provide management, scalability, and user experience. Further, it provides a tool to the developers to increase customization.