Research And Development Of Data Pre-Fetching Under Distributed Data Base

Distributed Data Base technology is a combination of Distributed technology and Data Base technology. It features the distribution of data and the coordination among Data Bases. The system should emphasize on the autonomy of the knots but not the central control. Also, it should keep the data distribution transparent, allowing not considering how the data are distributed when composing a program. However, the sharp increase of the network users and the intrinsic delay of network itself, make the computer network more and more crowded, and lack of good insurance of visiting speed and liability of Data Base. Searching manually becomes a bottleneck of the management of Distributed Data Base. Nowadays, when visiting a remote Data Base, various programs send request first to the server to get information. Pre-Fetching technology pre-fetch each result according to the current request, predict possible requests, and fetch those predicted to the local DataSet. Thus, users only need to download these contents from local DataSet when they do want to visit, greatly decreasing the visit delay.In this paper, we study and propose a Data Pre-Fetching System under Distributed Data Base with the help of Agent technology and a Data Mining method. Its main functions are as follows:(1) In the light of the lack of aptitude of the prevalent Data Mining system, we give a model of the data Mining system. It is based on Agent. And by the feature of Agent, it can realize the aptitude of the whole Mining process.(2) In order to let users predict, we makeup an interest-relative-rule knowledge base and give an interest-relative-rule-Mining algorithm to explore the users' interest-relative-rule from the users' historical behavior. Thus, we can predict the users' behavior, and improve the speed of the system when handling historical data.(3) In this paper, we design a Pre-Fetching System under distributed Data Base exampling Logistics. Because this system is based on Agent technology, take advantage of the data Mining method and fully consider the current interest of...