Processing search queries in a distributed environment
Frederick Knabe, Daniel Tunkelang · 2004
Endeca's approach to processing search queries in a distributed computing environment is predicated on concerns of correctness, scalability, and flexibility in deployment. Using a master-slave architecture, we are able to support classic search as well as more advanced features. We avoid bottlenecking the master with excessive computation or communication by limiting the information from the slaves in the expected case.