Monday, February 25, 2013

Enterprise Search functionality in Sharepoint 2013

Search in SharePoint 2013

The first and foremost change to search in SharePoint 2013 is there is only one search engine.

The idea that you use the FAST engine for content and the SharePoint engine for people has been completely eliminated. In SharePoint 2013 there is a brand new search core that combines the best qualities and functionality of both SharePoint search and FAST search. The powerful indexing, linguistics, extraction, and query expressiveness are now evident throughout the platform. With this change comes another dramatic change in overall search architecture. There is now only one query language across all of SharePoint 2013, which combines the best features of FQL (FAST Query Language),  and KQL (Keyword Query Language).

2.      Content Crawling

Anonymous Crawl

This adds a component to the crawling feature to support anonymous crawling. Prior to this change, all crawls needed to be associated to a user account or leverage other methods of authentication.

Continuous crawl

In SharePoint Server 2013, you can configure crawl schedules for SharePoint content sources so that crawls are performed continuously. Setting this option eliminates the need to schedule incremental crawls and automatically starts crawls as necessary to keep the search index fresh. Administrators should still configure full crawls as necessary.

In SP2010, we have two types of crawls; Full or Incremental Crawl. One of the limitations of Full and Incremental Crawls in SP2010 is that they cannot run in parallel, i.e. if a full or incremental crawl is in progress, the admin cannot kick-off another crawl on that content source. This forces a first-in-first-out approach to how items are indexed.

Moreover, some types of changes result in extended run times; such as script based permission changes, or moving a folder, or changing fields in a content type. Incremental crawls don’t remove “deletes”, so ghost documents are still returned as hits, after deletion, until the next full crawl.

SharePoint 2013 will introduce the concept of “Continuous Crawl”. It doesn’t need scheduling. The underlying architecture is designed to ensure consistent freshness by running in parallel. Right now, if a Full or Incremental crawl is slow, everything else awaits its completion. It’s a sequential crawl. Behind the scenes, continuous crawl selection results in the kick-off of a crawl every 15minutes (this wait can be configured) regardless of whether the prior session has completed or not. This means a change that is made immediately after a deep and wide-ranging change doesn’t need to ‘wait’ behind it. New changes will continue to be processed in parallel as a deep policy change is being worked on by another continuous crawl session.

 Note that Continuous crawl will increase load incrementally on the SharePoint server since it inherently can run parallel multiple sessions simultaneously. If needed, we can tune this through ‘Crawl Impact Rule’ settings (which exist today in SP2010) which controls the maximum number of simultaneous requests that can be made to a host (default is 12 threads, but is configurable).

Host distribution rules removed

In SharePoint Server 2010, host distribution rules are used to associate a host with a specific crawl database. Because of changes in the search system architecture, SharePoint Server 2013 does not use host distribution rules. Instead, Search service application administrators can determine whether the crawl database should be rebalanced by monitoring the Databases view in the crawl log.

Removing items from the search index

In SharePoint Server 2010, Search service application administrators could remove items from the search index by using Search Result Removal. In SharePoint Server 2013, you can remove items from the search index only by using the crawl logs.


Visual metadata extraction

Support for extraction of metadata using high-performance format handlers that allow for extraction of titles, authors and dates from hTML, DOcX, PPTX, TXT, Image, XML, and PDF formats. These format handlers look at font type and size, text alignment, capitalization, and other visual cues that  we ourselves will  generally use to determine the title and author of a document.

Company names entity extraction

First introduced in SharePoint 2010, company name entity extraction has undergone a serious facelift. Instead of extraction dictionaries being managed in XML files on the file system like in 2010, SharePoint 2013 now managed inclusion and exclusion dictionaries from within the term store of the Managed Metadata Service. This greatly simplifies the management and extension of this capability.

Result Sources

Formerly known as scopes in SharePoint 2010, the result sources tool in SharePoint 2013 now combines the 2010 concepts of scopes and federation into a new, powerful tool.

One of the most significant features within result sources is the support for remote SharePoint index

federation. While simple on the surface, this functionality fills a serious gap that existed in the overall scalability of SharePoint 2010. FAST and SharePoint were criticized in the marketplace for not having a global systems architecture. The approach was to tell users to centrally index all content in a large central farm, if the latency allowed. 

Remote SharePoint indexing addresses this problem by allowing federation with interleaving between local and remote SharePoint indices. This gives SharePoint 2013 a true global architecture solution that can be redundantly meshed to provide a scalable, fault tolerant architecture.
4.      Analytics Processing Components


Used in both feeding and query processes:
 Feeding: receives processed items from the content processing component and writes those items to index files
Query: receives queries from the query processing component and provides results sets in return

It also physically moves around indexed content when the index architecture is changed by the Search Administration Component 

FAST architecture is gone, but behind the screens, Columns are now referred to as "partitions" and rows are referred to as "replicas"

"Index partitions" and "index replicas" have the same conceptual behaviors as rows and columns. "Index partitions" allow you to index more content; "Index replicas" allow you to provide redundancy for your queries

6.      Query Processing Component

·         Performs linguistic processing at query time

·         Word breaking, stemming, query spellchecking, thesaurus

·         Analyzes and processes search queries and results.

·         When the component receives a query from the search front-end, it analyzes and processes the query to attempt to optimize precision, recall and relevancy. The processed query is then submitted to the index component(s).

·         As part of this it also decides which query rules are applicable, which index to send the query to, and whether to do any pre- or post-processing of the query

 The index component returns a result set based on the processed query back to the query processing component, which in turn processes that result set before sending it back to the search front-end.


Thumbnail Preview in FAST Search for SP 2010

In SharePoint 2013:

·         The new Office Web Apps is the engine for thumbnail previews in SharePoint 2013

·         The BIG WIN HERE – you can now browse through the entire document in the preview

·         See all pages, see animations, zoom in, scroll through the entire document

·         The point of this is to allow users to find the exact item they’re looking for right in search results – no more clicking a result, hitting the back button, and on and on until they find the one they’re looking for

·         It also addresses the two major shortcomings of thumbnails in SharePoint 2010:

·         It could only be used with FAST Search

·         It did NOT work with claims authentication 

·         Since there’s only one search engine in SharePoint 2013 you get document previews out of the box

·         In a different twist, previews only work with claims authentication – it will not work with classic Windows authentication

Please refer below links to know more about these types of authentication.

Search results will popup a "preview" window which includes an Office Web Apps document preview, but also includes some social media links, like "follow" the author of this document.

8.      Query Suggestions

It improves on the experience in SharePoint 2010 as follows:

·         Your personal SharePoint activity factors into the query suggestions, i.e. you have a personal query log

·         It includes weighting based on sites that you have previously visited

·         It uses the most frequent queries across all users that “match” the search terms

·         The behavior of the query suggestions turns into more of a “browse and find” kind of experience

·         You can also add inclusion and exclusion lists for suggestions via the SSA admin pages

·         When entering a query you will see two types of suggestions:

·         A list of items you have clicked on before from your personal query log

·         A list of items that others are typing for their queries

·         When you get query results back, you will get another set of suggestions

·         They are a list of links that you have clicked through at least twice before and match your search criteria

There are two different modes for the refiner web part: search results and faceted navigation.

·         With search results the refinement data works essentially the same as SharePoint 2010

·         With faceted navigation it uses a term from the term store to filter what kind of data should be displayed.

·         Refinement is different with SharePoint SharePoint 2013 in that you can define display templates to use for rendering different kinds of refinements.In SharePoint 2010 you had to write your own custom refiner.

 Faceted Navigations:

·         With Faceted Navigation, it is used in conjunction with term sets that are used for navigation

·         With each term you select which managed properties should be used as refiners with that term

·         Within the managed property you need to configure it as “Refinable”

·         Example:

·         You have term store terms Camera and Laptop

·         You have managed properties Megapixel Count, Color and Manufacturer

·         For Camera term, you add refiners for Megapixel Count and Manufacturer

·         For Laptop term you add refiners for Color and Manufacturer

 Note: Refiners no longer provide hit counts; it was felt that they were not used much so they have been removed.  

The following section provides details about the deprecated features in FAST Search Server 2010 for SharePoint.

FAST Search database connector

Description: The FAST Search database connector is not supported in SharePoint 2013.

Reason for change: The connector framework for SharePoint 2013 is combined with the BCS framework and the Business Data Catalog connectors.

Migration path: Replace the FAST Search database connector with the Business Data Catalog-based indexing connectors in the BCS framework.

FAST Search Lotus Notes connector

Description: The FAST Search Lotus Notes connector is not supported in SharePoint 2013.

The Lotus Notes indexing connector (BCS framework) provides similar functionality as the FAST Search Lotus Notes connector. The FAST Search Lotus Notes connector supports the Lotus Notes security model. This includes Lotus Notes roles, and lets you crawl Lotus Notes databases as attachments.

Reason for change: The connector framework for SharePoint 2013 is combined with the BCS framework and the Business Data Catalog connectors.

Migration path: Replace the FAST Search Lotus Notes connector with the Lotus Notes indexing connector, or with a third-party connector.

FAST Search web crawler

Description: The FAST Search web crawler is not supported in SharePoint 2013.The SharePoint 2013 crawler provides similar functionality to the FAST Search web crawler

Reason for change: The crawler capabilities are merged into one crawler implementation for consistency and ease of use.

Migration path: Use the standard SharePoint 2013 crawler. The following table explains the differences between the FAST Search web crawler and the SharePoint 2013 Preview crawler, and provides details about migration. 

Find similar results

Description: The Find similar results feature is not available in SharePoint 2013. The Find similar results feature is supported in FAST Search Server 2010 for SharePoint to search for results that resemble results that you have already retrieved.

Reason for change: The Find similar results feature is available only within the query integration interfaces, and it does not consistently provide good results in many scenarios.

Migration path: There is no migration path available.



Description: The search anti-phrasing feature in FAST Search Server 2010 for SharePoint is not supported in SharePoint 2013.

Anti-phrasing removes phrases that do not have to be indexed from queries, such as “who is”, “what is”, or “how do I”. These anti-phrases are listed in a static dictionary that the user cannot edit.

In SharePoint 2013, such phrases are not removed from the query. Instead, all query terms are evaluated when you search the index.

Reason for change: The FAST Search Server 2010 for SharePoint feature has limited usage due to the limited number of customization options.

Migration path: None.


Substring search

Description: The substring search feature was removed in SharePoint 2013.

In FAST Search Server 2010 for SharePoint, substring search (N-gram indexing) can be used in addition to the statistical tokenizer in East Asian languages. Substring search can be useful for cases in which the normal tokenization is ambiguous, such as for product names and other concepts that are not part of the statistical tokenizer.

Reason for change: The feature has limited usage, and has very extensive hard disk requirements for the index.

Migration path: None


Number of custom entity extractors

Description: In SharePoint 2013, the number of custom entity extractors that you can define is limited to 12.

In FAST Search Server 2010 for SharePoint Service Pack 1 (SP1), you can define an unlimited number of custom extractors. You can use custom entity extractors to populate refiners on the search result page.

There are 12 predefined custom entity extractors in SharePoint 2013:

·         Five whole-word case-insensitive extractors

·         Five word-part case-insensitive extractors

·         One whole-word case-sensitive extractor

·         One word-part case-sensitive extractor


Reason for change: By using a predefined set of custom entity extractors, the content processing architecture is more simple and easier to use.

Migration path: Use the predefined set of custom entity extractors.


Supported document formats

Description: SharePoint 2013 no longer supports rarely used and older document formats that are supported in FAST Search Server 2010 for SharePoint by enabling the Advanced Filter Pack. Both the ULS logs and the crawl log indicate the items that were not crawled.

In SharePoint 2013, the set of supported formats that are enabled by default is extended, and the quality of document parsing for these formats has improved.

Reason for change: The file formats for indexing are older formats and are no longer supported.

Migration path: You can work with partners to create IFilter-based versions of the file formats that can no longer be indexed.

Custom XML item processing

Description: FAST Search Server 2010 for SharePoint includes a custom XML item processing feature as part of the content processing pipeline. Custom XML item processing is not supported in SharePoint 2013.

Reason for change: In SharePoint 2013, the content processing architecture has changed. Custom XML item processing was removed and we recommend that you implement a mapping functionality outside SharePoint.

Migration path: Custom XML item processing can be performed outside the content processing pipeline, for example by mapping XML content to a SharePoint list, or to a database table

11.  Key Points about SharePoint 2013 Search


·         Office Web Apps is no longer a service application

·         Seperated to own product

·         Web Analytics is no longer service application

·         Analyses and reporting process incorporated to search service application

·         Overall SharePoint 2013 requires more resources than 2010

·         SharePoint crawler will support "anonymous" authentication for crawling website

·         Creating crawled and managed properties still require a FULL crawl

No comments:

Post a Comment