froscon2008 - 1.1
FrOSCon
Free and Open Source Software Conference
Referenten | |
---|---|
Ralf Schwoebel |
Programm | |
---|---|
Tag | Day 2 (2008-08-24) |
Raum | HS3 |
Beginn | 10:00 |
Dauer | 01:00 |
Info | |
ID | 143 |
Veranstaltungstyp | lecture |
Track | Web |
Sprache der Veranstaltung | en |
Feedback | |
---|---|
Haben Sie diese Veranstaltung besucht? Feedback abgeben |
(Web) Search on Linux driven sites
High speed search with SPHINX or MNOGO for huge amounts of data
This lecture demonstrates the installation, configuration and usage of search and index servers on Linux for speeding up web searches on sites with 10 Million (or more) entries / pages. Open Source packages used: SPHINX as a MySQL (among others) search engine and MNOGO as a full scale webspider and search engine.
Fast growing websites gather huge amounts of information and large e-commerce sites operate with catalogs containing millions of entries.
For "on the fly" requests of related products, a precise and detailed search of millions of documents or a simple user name query speedy application frameworks are crucial for success.
Faster web pages convert better and create satisfied customers!
This track will give a practical insight into the installation and use of
- Sphinx (http://www.sphinxsearch.com/)
the free open-source SQL full-text search engine, which is used by our company to search more than 10 million database entries under load daily. Sphinx can be run on top of MySQL databases. The simple and powerful PHP API allows an easy integration into PHP pages. An example will be shown.
- mnoGoSearch for UNIX (http://www.mnogosearch.org/)
a free software covered by the GNU General Public License. It consists of a command line indexer and a search program which can be run under Apache Web Server, or any other HTTP server supporting CGI interface. mnoGoSearch is a simple alternative to Sphinx, if you need to keep millions of URLs of several domains indexed and searched via a web interface.
Both softwares are used by Tradebit AG to run a variety of sites for years. Problems, instability or tricks for a continuous online service on both packages will be discussed at the end of the lecture.