Index Data is pleased to announce a major upgrade of Zebra, its Open Source database server and indexing engine. Zebra 2.0 represents an extensive upgrade that makes index profiling much easier, supports increased tuning of search results, incorporates XML technology into core functionality, and dramatically increases performance speed.
Some of the highlights of the many improvements of Zebra 2.0 over the 1.3 version are:
- A 64-bit based index structure
- Elimination of the 2GB limit on register file size
- New on-disk format providing increased stability and faster indexing and retrieval
- New record filter using XSLT transformations to drive both indexing and retrieval
- Improved logging and analysis of external traffic
- Revised and expanded documentation
Zebra 2.0 replaces the previous versions' tight coupling to the Z39.50 BIB-1 attribute set with a new XML friendliness, making Zebra easy to use for such XML-based formats as Dublin Core, MODS, METS, MARCXML, OAI-PMH, RSS, etc. The software's new plug-in architecture allows the skilled user to write his or her own record indexing and retrieval filters as loadable modules.
The performance enhancements incorporated into version 2.0 mean that Zebra can now index and search even faster than version 1.3. In a test of Zebra 2.0, the software was able to build a 31 GB database of very large records in four elapsed hours on a 1800 GHz Dual AMD box. It processed an average of 2.2 MB of data per second.
Zebra 2.0 offers more precise logging of external traffic, access and indexing, and log messages are now printed in a style similar to Apache server logs. These enhancements greatly improve users ability to mine the log data for patterns of use.
Overview of Zebra
Adam Dickmeiss and Sebastian Hammer, the co-founders of Index Data, began developing Zebra in the mid 1990s. It was originally designed as a native Z39.50-based indexer of MARC bibliographic records, but has evolved into an XML-based text indexing engine and database server that can handle full text as well as any form of structured metadata. It is used by hundreds of libraries and library vendors around the world for databases with upwards of 70 million records. For more information on Zebra, visit www.indexdata.com/zebra.