#SPC09 Day 2 – An Intense Day on Enterprise Search

21. October 2009
Comments (0)

Today was a big search day at the SharePoint 2009 conference in Las Vegas. Five sessions packed with information on new and improved features in SharePoint 2010 Enterprise Search. I attended all the sessions but one and have learned a great deal on the new search experience since my last posting on Sunday.

My general impression on search from the sessions is very good – it seems like Microsoft has really done a good job listening to customers, partners and the community and fixed some major pain points identified with SharePoint 2007 Search. On top of that we are also going to see some new and very useful features. But before I start enumerating any features it is helpful to know that Microsoft will offer three Enterprise search products:

  • Search Server 2010 Express. Free Enterprise Search engine with certain restrictions. Don’t know the details yet but my guess is the limitation will be one server deployment only and a few missing features like Business Connectivity Services (BCS) and People Search.
  • SharePoint Server 2010. Ships with the full version of the Enterprise Search engine including BCS and People Search. Suitable for customers without extreme scalability and extensibility requirements. Scales to approx. 100 million documents.
  • FAST Search Server 2010 for SharePoint. For customers with extreme scalability, functionality and extensibility requirements. Scales to billions of documents and thousands of concurrent queries.

The following list of new and improved features applies to the first two products:

End-User Improvements

  • Improved relevance. More parameters included in score calculation. An important new parameter is click-through rate on search results also known as popularity ranking, i.e. popular results tend to bubble upwards. Also support for customizing the relevance algorithm.
  • Enhanced query syntax. Support for Boolean operators like AND, OR, NOT. Also support for the range operators <, >, <=, >= for search numeric ranges or date ranges.
  • Wildcard search. Now possible to search for partial words using the wildcard character *. E.g. Micro* author:bill*
  • Phonetic and nickname search. Useful in people search to match similar names with different spelling. E.g. queries for people named “Jacob” also returns people named “Jakob”.
  • Faceted search with support for refining search results on any managed property configured to support refinement. Offers a very intuitive way to filter results on metadata – will make advanced search less relevant to most users.
  • Query suggestions as you type. Mined from query log.
  • Improved did you mean. Support for more languages.
  • View in browser. Users can from the search results now view Microsoft Office documents directly in the Web browser without having the Office client installed. Requires the new Office Web Apps on the server.
  • Desktop search integration.

Administration Improvements

  • Improved admin dashboard.
  • Common admin experience across all three search products.
  • PowerShell. All administration features are scriptable via PowerShell.
  • Search Reporting. Extensible search analytics reporting.

Scalability and Reliability Improvements

  • New search engine architecture. Provides better better scale-out and scale-up options.
  • Multiple crawlers (formerly known as index servers) can now work in parallel building an index. Turns the role into a redundant role and enables greater crawl speeds. Microsoft presenter reported 175 docs/secs with four crawl servers.
  • Partial indexes. The full index is no longer required to reside on each query server. It can now be split over multiple query servers for greater query performance. Can also be mirrored for increased availability.

Extensibility Improvements

  • Open Web parts. All the search Web parts are now public and open for inheritance. Makes it possible to add more functionality with few lines of code.
  • Common UI framework across all search products including FAST.
  • Common APIs across all search products including FAST.
  • New connector framework for indexing LOB systems. Essentially a managed code replacement for protocol handlers. C/C++ protocol handlers are, however, still supported.

Other Improvements

  • Improved Wordbreaker with support for more languages.

I can not say much about the improvements to FAST Search Server 2010 for SharePoint other than it can do all that the standard SharePoint 2010 search engine can do plus a lot more.

SharePoint 2010