FEDERATED SEARCH
OVERVIEW
The Federated Search service allows a query to be sent to a large set of disparate data providers and then collecting the results generated by each and presenting them back to the user. The returned search results are de-duplicated, ranked, etc. before returned to the user. This allows a user to submit a query from one place using one syntax and retrieve relevant data from many sources across DoD. This approach leverages existing data sources and production processes.
Federated Search is implemented as a set of cooperating web services. These services talk to each other using a common specification. The specification defines how a query and the results from that query are communicated. It describes not only the meaning, but also the format of the data that is exchanged between the services.
The Defense Discovery Metadata Specification (DDMS) is used by the Federated Search service to represent the concepts of a query as well as the resource result records, called meta cards, generated by a search result. Outgoing queries are matched against the resource meta cards by data providers to generate search results. It is the DDMS that ties the queries to the results and is used to express a common vocabulary.
The domain of the Federated Search service is limited to the provider sites made available by the sponsoring organizations for the DoD enterprise. The Federated Search service does not provide visibility or access to private provider sites that do not participate in the Federated Search service.
The Federated Search service includes the following functions:
- Distributes queries to selected content providers
- Aggregates returned results
- Deduplicates results
- Ranks results
- Returns results to query source
- Controls visibility to content based on security profiles utilizing NCES SOA Security Services
The Federated Search service interface includes the following characteristics:
- Supports Representational State Transfer (REST) and the Simple Object Access Protocol (SOAP) interfaces
- Makes use of industry standards where practical (XQuery, ATOM)
- Minimizes cost of implementation by content providers and consumers
- Supports relevance ranking
