OpenSearch in Alfresco 2.0 Federates the Distributed Organization

Alfresco's Chief Architect Dave Caruana’s blog post on Alfresco’s support for Open Search in our version 2.0 should not go unnoticed. Dave's latest contribution to Alfresco is a big deal and I think it is a first in the Enterprise Content Management space. OpenSearch was originally created by A9 to provide a mechanism to aggregate search results from multiple search sources. Now it is supported by literally hundreds of search engines. Alfresco is one of the latest pieces of software to support OpenSearch.

Alfresco supports OpenSearch both as a client and as a server. This means that you can include one or more Alfresco repositories as well as the internet or any other search engine in your web browser, a portlet or other type of search tool. Firefox 2.0 and IE7 support OpenSearch from their built-in search tools and it is now easy to add Alfresco searches. In addition, you can search multiple Alfresco repositories and the internet from the Alfresco web client. This is a powerful tool to bring content into a repository and to aid in collaboration. We have been able to use both tools in customer implementations that include multiple Alfresco repositories with blogs, wikis and external search engines.

This brings a whole new definition to Federated Repositories. Alfresco can now join federations of search engines and collections of repositories. Existing tools like IBM’s Venetica and EMC’s AskOnce use proprietary connector technology that relies on centralized topologies of integration. Because these searches rely on proprietary interfaces, what repositories and sources are supported depends on the vendors or after-market suppliers to provide the connectors. If these engines wanted to include the list of information source provided above, then they will eventually have to support OpenSearch.

The Alfresco approach is different in that it uses a standard protocol and allowing departments and individuals to configure how they want to federate their searches. It allows for a loosely coupled topology of repositories that can grow and fuse repositories as business needs require. Departments can try Alfresco at their leisure and then merge results through federated search, even if the collaborating organizations are outside the enterprise. Also, searches do not need to be limited to what Alfresco supports. The user can include searches to information sources that are important to the task at hand. This is a leap away from the repository-centric view where all activities revolve around the repository to a holistic view of information in which the repository plays a supporting role.

The OpenSearch really a collection of technologies based upon relatively simple, standardized protocols. http://en.wikipedia.org/wiki/OpenSearch The search engines are described using XML that is extensible and therefore expandable to support more sophisticated ECM types of searches. The search itself is invoked using a URL and therefore fits the web and REST model very neatly. The search is usually described in a template form, but is extensible to add additional metadata types of searches and is designed to include metadata. Page results are returned in the form of ATOM, RSS or HTML and well suited to work within a browser or to be aggregated in an aggregation tool. We have been able to use an open source Java aggregator as part of our web client implementation. (That’s the beauty of open source - we don’t have to reinvent it every time.)
Please give it a try. You can access the latest version of 2.0 at http://www.alfresco.com/products/docs/releases/2.0/

Comments

[David Caruana]

I can't take all the credit. Gavin Cornwell implemented the OpenSearch user interface.

Dave