| 摘要 |
A method and apparatus for applying a query to a database that is distributed into first portions. The query is first partially performed by terminals each containing one of the first portions. The results of the partial queries are recombined in such a way as to minimize the traffic among the terminals connected across networks. Paths of an edge-labeled tree representing the database connecting the first portions are expanded into subpaths corresponding to all the states of an automaton modeling the query. The database is distributed into second portions based on the automaton. Non-existing subpaths are ignored based on the automaton and which of the second portions the paths exit. Thus, the data transferred among first portions are minimized and the query process is optimized.
|
