September 11, 2008, 5:08 PM — A U.K. startup called Simplexo has launched an open source enterprise search engine, joining a small but growing group of vendors hoping to crack the crowded technology space through that development model.
Simplexo -- which joins open source offerings from the likes of Tesuji and Lemur Consulting, as well as scores of proprietary tools -- uses a hybrid approach to searching content. It places unstructured data, such as Web pages and Word documents, into an index, but directly queries structured data sources like databases and enterprise applications, according to a report by Rob Hailstone of the U.K. analyst firm Butler Group.
This makes for a smaller index, requiring less work to keep the index up-to-date, the report said: "Because databases and applications are queried in real time, the results from these will always be current."
"The most likely performance constraint with this architecture will be slow response from external applications, although this is mitigated by the fact that all sources are queried in parallel rather than being serialized," Hailstone added.
The product "has been developed to Ministry of Defence security levels for your peace of mind," according to Simplexo's site.
The source code is available under the GPL v2 license. As is common for open source companies, Simplexo will look to make money selling optional support.
The vendor is developing and testing the software in-house, while "community-recommended additions and modifications that fit into the overall product development strategy will be incorporated into the Simplexo development and release schedule and made available with a new, supported version," Hailstone's report said.
One longtime observer of the enterprise search space had a measured reaction to Simplexo's announcement, as well as the notion of open-source search in general.
"I am neutral on open source search solutions," said analyst Stephen Arnold in a blog post Thursday. "If you have the technical resources, open source can deliver excellent results. If you are not comfortable with open source, then you may be better served by running a try-before-you-buy analysis and then a bake-off. Let your data collection guide you."