Studies On Database Notification Engine Based On Rete Algorithm

Nowadays we are in an information-booming era. Many activities, such as trading in stock market, sports competition activities, production management, company operation, advertising activities of the shopping mall, produce large quantity of information that we pay close attention to. The information, however, are mostly stored and managed by database systems. With the addition of new information, the data in databases are updated constantly every-day. To avoid the unnecessary economic losses, the users and customers of the databases often hope to be noticed while their concerned data was changed. On the other hand, they often hope to know the states of certain data in databases at regular interval in order to timely adjust or improve the operating strategy or management schema.Database Notification Services are exactly software that can meet the above-mentioned demands. At present, Database Management System providers such as Sybase, Oracle, Microsoft, IBM, etc. all have put out the developing platforms for notification service. However, their solutions exist some insufficiencies. Firstly, every provider only offers separate service for customers. Every platform only supports a specific type of DBMS, and can't support others. Many heterogeneous DBMSs, however, are used in some enterprises simultaneous. When a nofication service uses data from two or more heterogeneous DBSs, their solutions will not work. Secondly, most notification services use SQL to code the customers' subscription and match subscriptions with the events through SQL query. Nevertheless, SQL is limited in the recursive query that is used in notification services sometimes. Lastly, changing anything by SQL coding is very inconvient.The Natabase Notice Engine is the core of Database Notification Service.This paper adopts some research results from "Data Integratation and Pattern Match on Hererogeneous Database systems Based on Inteligent Agent Technology", which is funded by the Youth Foundation of Shanghai College Committee, and design and realize a prototype for the Database Notification Engine. This system is totally developed in Java language. The following research achievements have been achieved:1) Offer a new way to realize Database Notification Service. Namely, consumers' database notification tactics are coded by production rules instead of SQL. Production rules are easier for consumers to understand and accept, and production rules can be modified more easily too. Furthermore, by using production rules, we resolve both database recursive queries and complicated reasoning on which SQL can't do any help.2) Design and realize a visual rule editor. The rule editor make it more convenient for consumers and users to define a database notification tactics;3) Define and realize the interface between the rule engine and the database. The interface and the rule engine are encapsulated as Database Patterns-matching Engine. Rete algorithm is used in the Engine as the patterns-matching algorithm.4) Realize notification services among heterogeneous database systems by using JDBC.5) Design and implement Database Notification Engine. We realize a prototype system at last, which can send the notice to consumers through SMS and E-mail.