Open Source eDiscovery Solutions

Search Technologies is an independent provider of implementation services for eDiscovery and other search-centric solutions. The company employs 75 staff (as of April 2010), primarily in the USA and the UK, and serves more than 100 government and corporate customers.

Search Technologies maintains an open source services practice which builds eDiscovery applications based on the Apache Solr Lucene open source search engine. Solr Lucene is probably the most widely deployed search engine Worldwide, and it has reached a level of stability and reliability equivalent to the best commercially licensed alternatives. Importantly, there is now an extensive community working with Solr Lucene, and new functions and add-on capabilities are quickly appearing and reaching commercial-equivalent reliability.

Search Technologies provides complete, project managed eDiscovery solutions based on Apache open source products; from requirements review and architectural design, through implementation and project management, to ongoing maintenance and support under a commercial-grade service level agreement. With the addition of professional services to customize the application to the specific needs of client, open source now provides a highly cost effective option for serious eDiscovery solutions. Further, where client needs are quite specific, open source affords great flexibility, and being open source-based, the client is not subject to vendor tie in.

Clients can freely choose to take ongoing support in-house, or to contract with competent 3rd parties as necessary. eDiscovery solutions can be built using the following Apache Lucene open source components.

• Lucene: A core indexing and search API

• Solr: A server packaging for Lucene providing a wide range of enterprise search functions and a convenient RESTful/xml interface

• Tika: A document filtering / content extraction utility

• Mahout: A machine learning add-on for Lucene supporting clustering, collaborative filtering, patternset mining, random decision forests and support vector machine implementation.

• Nutch: A crawler for use inside or outside the firewall In some specific cases,

For some eDiscovery applications it may be pragmatic to use a mixture of Apache open source components and one or more licensed products. Search Technologies provide independent advice in this matter, on a case by case basis.