Indexing 17: distributed search

Victor Lavrenko
Victor Lavrenko
7 هزار بار بازدید - 11 سال پیش - Instead of using MapReduce to
Instead of using MapReduce to construct a single index, we can distribute portions of the index across a cluster of machines. We can then send a query to all the machines, receive partial ranked lists and then combine them into one list that would be returned to the user. This is known as distributed search, federated search or meta-search.
11 سال پیش در تاریخ 1392/10/30 منتشر شده است.
7,067 بـار بازدید شده
... بیشتر