The federation engine should allow clients to search over the federated content from multiple repository sources. This may be possible through search indexes (and crawling) or by pushing down queries to the connectors and integrating (union-ing) the results.