This document summarizes various techniques for scalable continual top-k keyword search in relational databases. There are two main approaches: schema-based and graph-based. Schema-based methods generate candidate networks from the database schema and evaluate them. Graph-based methods represent the database as a graph and use techniques like bidirectional expansion. Top-k keyword search finds the highest scoring k results instead of all results. Methods like the Global Pipeline algorithm and Skyline-Sweeping algorithm efficiently process top-k queries over multiple candidate networks. Techniques for updating results with database changes include maintaining an initial top-k and recalculating scores. Lattice-based methods share computational costs for keyword search in data streams.