Sesat > Features > Technical Feature Matrix

Feature Matrix

We keep an up to date and detailed Kernel Feature list here.

Feature Description
Installation and environment  
Linux flavours All of course, but tested on CentOS, RedHat, Ubuntu and Debian. Works well on local Windows installations (not recommended for heavy production sites)
Application Servers Supported Tomcat 6.0+, JBoss 4.2.0+
Apache front-end proxying supported with mod_cache and mod_jk Yes
Runs on same server as client application No need for separate server(s). SESAT is deployed directly on the Java Application Server
Supports Java 1.6 Yes
Multiple Sites in the same JVM Yes, with properties and templates inheritance for (sub)sites (e.g. vg.sesam.no can inherit properties and settings from sesam.no)
Easy configuration of each defined site/application Yes, currently by XML-files, but with GUI-administration module under development.
Scalability Scales linearly with number of servers using HW load balancing with sticky sessions.
Automated Deployment Deployment scripts included for deployment of both SESAT and client applications.
Source Code  
Full access to Source Code yes, from Git
Permissions to apply own patches Yes
Special functionality built upon request Yes, contribute it yourself or persuade the developers.
Site GUI Development  
Search Portal Framework Yes. SESAT is both federating search middleware and a search application framework.
Development Environment Anything you want. We recommend both NetBeans and Eclipse.
Development OS Linux, Mac OS X and Windows
Client application building Maven2-based generation of .war deployment files
Content formats supported HTML, XHTML, XML, VCF, VML, RSS, etc.
Site inheritance SESAT allows for sites to inherit layout and configuration from parent sites.
Example templates included Yes, extensive stack of templates based on several real-life sites. (Velocity or JSP)
Easy development environment setup Yes. Supports ssh-tunnels to data sources.
Search result access method The Search Result Object is directly available from the template system.
Normalised search result The Search Result Object presents the result in a normalised and uniform way, independent of the underlying index technology.
Templating System Velocity and JavaServerPages are standard, support for adding any other Java-based templating systems.
Ajax-support Search Result Object is connected to user sessionID to provide asynchronous information access.
Support for Java Faces Not as standard yet, but can easily be implemented by client. Near-future versions of SESAT will have working examples of Java Faces.
Full flexibility for templating framework Yes.
Query Analysis  
Configurable business logic Yes, through the rules engine and administrative tools.
Rules-engine based Query Analysis Yes, with rules managed by own module in SESAT ADM
Automatic search-phrase highlighting Yes, for FAST-indexes. Highlightning independent of index type will be released in the near future.
Specification and administration of trigger-words Yes, triggers may change ranking order of individual indexes in the search result. For example, "fact oslo" may boost content from a Wikipedia index before content from a global crawled web index.
Administrative back-end for rules engine Yes, SESAT ADM Text Analysis providing automatic detection of full-names, geographical names etc.
Search term categorisation based on lookup-lists yes
Phrase-search support yes
Standard Search Operator support yes, all functionality is inherited from the underlying index motor
Easter-eggs support yes, of course!
In-built calculator Yes, queries can be mathematically evaluated
Result Handling  
Result Object Comprehensive, easily accessible object model
Simple Flat Result List Search result represented as java object
Result persistence Yes, attached to user SessionID, configurable duration
Handle both generic and specific index results Yes - user configurable and extendable
Enrichments Support  
Enrichments ranking based on rules yes
Easy creation and definition of enrichments Yes, you can use any search command as an enrichment that will be displayed and ranked according to initial search analysis.
Unlimited number of enrichments Yes
Ad Systems Support  
Yahoo PPC yes
AdMomentum yes
Google AdWords yes
Easy to add new Ad Systems Easy for user to implement Ad Systems by creating new, simple search commands
Data Stores Support  
Default Data Index Support Default Search Command implementations for Fast-4 Simple search, Fast-4 Advanced search, Fast-5 ESP search, generic XML search, Web Services seach, Blending (FAST) search, Clustering (FAST) search, Correcting (FAST) search, Mathematical/Calculator search, Overture PPC search, PicSearch search, Platefood PPC search, Sensis search, Solr (Lucene) search, Stock market search, Yahoo IDP search, Yahoo Media search, and Youtube search.
Support for external data stores Yes, through HTTP, WebServices, JDBC. Easy to add own data sources using the SPIs.
Roadmap Data Source/Index Support Lucene, Google, SharePoint, other sources we may need to search in
Reference Search Command implementations Blinkx Video search, Blocket search, Finn search, GeoData Map search, HittaMap search, Hitta search, HittaWeather search, Prisjakt search, Storm Weather Blinkx Video search, Tasteline search, TV-guide (FAST) search, Whitepages (FAST) search, Yellow Pages (FAST) search
Support for Custom Search Commands Easy to develop using inbuilt SPI
Core Application Extensibility  
Available SPIs (Service Provider Interfaces) for extending core functionality Yes
Full access to source code Yes
Independent release-cycles of minor kernel upgrades vs. client code Yes
Backwards-compability Complete Upgrade Guides kept up to date.
APIs      
WebServices/XML support Any search result is available through XML interfaces.
SOA-enabled interfaces Yes, through WebServices
Search Logging Support  
Logging of requests to underlying search engines Yes, everything can be logged with timestamp and result information (no. of hits)
Logging of client information Yes, user environment, browser characteristics
Search context logging Referer-domain, "skin" (client site from which the user is searching), navigational elements.
User navigation tracking Yes, using SESAT's Boomerang component, you can track user behaviour on your site, and also any access to external sites. Also, any asynchronous access activities can be logged by Boomerang.
Administrative tool for reading logs Yes, part of SESAT ADM
Automatic anonymisation of IP-address Yes, configurable part of the statistics parser.
System behaviour logging For each search, SESAT logs search execution times (on a per-index basis), zero-hitcount search results (on a per-index basis), enrichments, slow responding indexes
Monitoring/Surveillance Logging Support  
Slow index/non-responding index detection Yes, logged to own surveillance log
Personalisation      
Basic personalisation support Yes
Persistence storages supported MySql currently supported, but other databases available from JDBC
Security user session keys generated at each request
Sample code Yes, working example of personalisation provided
User credentials storage OpenLDAP
Specific Index Functionality: FAST  
Full navigator support Yes, through Result Object and template system
Full support for rank profiles Yes, the profile can be specified in the search command.
Full support for collections Yes
Robust Error-handling for common FAST-errors Yes, for example query matching failures/downtime
 © 2007-2009 Schibsted ASA
Contact us