摘要 |
A selection engine for a cooperative processing network, in which databases compute portions of searches and analyses which they may most efficiently compute. A data mining network may contain large scale databases, analytic engines which communicate with the databases, and other resources, each capable of performing or executing certain functions, such as statistical functions. When a user desires to run a specified report against the databases, those functions that can be computed locally in the database are trapped and computed therein, whereas other functions, such as advanced statistical functions, may be distributed to remote analytic engines or other resources in the network. A management module may coordinate the transmission and assembly of the data, including to order dependencies. The management module may contain a selection engine to allocate computations to databases or other resources as a default, to test for load balancing before choosing a compute site, or apply other criteria. |