What are some Search Servers out there?

Question

I'm looking to find alternatives to Solr from the Apache Software Foundation.

For those that don't know, Solr is an enterprise search server. A client application uses a web-services like interface to submit documents for indexing and also to perform search queries. Solr has other features built in like caching and replication. I believe it was originally started by CNet and then open-sourced.

I'm looking for other search servers out there that might be seen as the competition.

Solution 1

I wrote a long post about my experiences and features of all the engines I listed below but I scrapped it because formatting is a pita. But quite simply if you don't want to shell out money Solr/Lucene or Fast (now MSSE) is really about the best you can do.

Excluded because I have no experience of this product: Seamark,

Price High to Low

• Endeca,
• FredHopper,
• Mercado,
• Google Mini,
• Microsoft Search Server,
• Autonomy,
• Microsoft Search Server Express,
• Solr/Lucene

Speed Fast to Slow

• Google Mini/Endeca,
• FredHopper,
• Autonomy,
• Solr/MSS/MSSE

Features High to Low

• Endeca,
• FredHopper,
• Mercado,
• Solr,
• Autonomy,
• Lucene,
• MSS/MSSE,
• Google Mini

Extensibility High to Low

• Solr/Lucene,
• Endeca,
• FredHopper,
• Mercado,
• Autonomy,
• MSS/MSSE,
• Google Mini

Java API

• Endeca,
• FredHopper,
• Autonomy,
• Solr/Lucene

.NET API

• Endeca,
• Solr/Lucene,
• MSS/MSSE,
• Autonomy

XML API

• FredHopper,
• Mercado,
• Solr/Lucene,
• Autonomy,
• Google Mini (limited)

Faceted Search

• Endeca,
• FredHopper,
• Seamark,
• Solr

Natural Language Search

• Endeca,
• Fred Hopper,
• Solr,
• Mercado,
• MSS/MSSE,
• Autonomy,
• Google Mini

Document Crawling

• Endeca,
• Mercado,
• MSS/MSSE,
• Autonomy,
• Google Mini

ITL

• Endeca,
• FredHopper

Merchandizing/Content Spotlighting

• Endeca,
• FredHopper,
• Mercado

Distributed Search

• Endeca,
• FredHopper,
• Mercado,
• Solr/Lucene,
• Autonomy,
• Google Mini

Analytics

• Endeca

Platform x86 Windows

• Endeca,
• FredHopper,
• Mercado,
• MSS/MSSE,
• Solr/Lucene,
• Autonomy

Platform x64 Windows

• Endeca,
• FredHopper,
• Solr/Lucene

Platform x86 Unix Variants

• Endeca,
• FredHopper,
• Mercado,
• Solr/Lucene,
• Autonomy

Platform x64 Unix Variants

• Endeca,
• Solr/Lucene

Other

• Google Mini

Solution 2

I've worked peripherally with Endeca, and it seems like a feature packed, fast engine.

They seem especially well suited towards eCommerce site driving, with special tools for customizing content for up-sells and cross-sells. I also think they are going to try and move into the BI space, so it should be an interesting combination of the two capabilities if they pull it off seamlessly.

Solution 3

I used to work at Endeca so take this for what it is. I won't comment on competitors, but I do really believe it's a significantly better product than anything else out there.

Solr provides basic faceted search, which is "good enough" for many applications. In a lot of ways, it's a low-end Endeca.

Some of the things Endeca would give you over Solr:

• better language analysis (stemming, etc.)
• Joins and the ability to do many-to-many filtering
• support for querying via XQuery
• better support for managing dimensions / facets
• much better performance (both for queries and pushing data in)
• crawlers and data integration tools
• better admin tools

And, obviously, services and support.

That said, it's not cheap, so if you don't have a budget for it, Solr isn't a bad option.

Solution 4

Here are some more open source search engines

• Sphinx
• Xapian
• Zend_Search_Lucene

Here are some search related tools

• Forage The Forage PHP5 library is an easy to use interface to multiple back end search libraries. It provides a common interface while supporting unique features in each library by allowing back-ends to support specific features or not.

• Marjory Marjory is a webservice for indexing and searching for documents, utilizing a full-text search engine.

Solution 5

+1 for Lucene and FAST. Lucene has also been ported to .NET if you're interested in extending it.

http://incubator.apache.org/lucene.net/

View more solutions

Comments

[stephbu]

Its interesting that IBM used FAST for it's site search, not it's own products.

[zeridon]

Fredhopper is 64 bit on Linux/Unix also.

[MovGP0]

ElasticSearch seems to be missing here.

[MovGP0]

also: Microsoft Search Server and Google on Premise is dead now.