The Research Of Application Using Rule Engine In Business Logic Layer

Most enterprise Java application can be divided into three layers: View layer, Persistence Layer (database) and Business logic layer between those two layers. Usually we use Framework (View layer likes Struts, Spring and so on, Persistence Layer likes Hibernate as well as Entity Beans etc.) to deploy the application, this means you can ignore many bothersome details, and focus on business logics. But so far, there is no such criterion on how to build logic layer.Rule engine is a kind of inference engine, it can trace back to rule-based expert system. Rule-based expert system imitates humanity's inference way, uses exploratory method carries on inference, also uses terminology which humanity can understand to explain and to prove its inference conclusion. Using this kind of inference engine can separate business rules from the application codes, accept the data inputs, explain the business rules, and make decisions according to the rules.Via the research of rule engine's principle and its core module - Rete algorithm, this paper reveals rule engine's inference process, uses rule engine to build the business layer and proposes a better solution on how to build the business layer.Moreover, through an actual small order form system utilizes the rule engine framework to find how to use rule engine framework in the business logic layer of enterprise Java applications. This order form system uses Struts + Drools + Hibernate framework and uses two ways to implement the business logic layer, one use the rule engine framework and the other one does not. Finally, compares these two implementations with the performance test and test result as well as other comparison such as development difficulty comparison and maintenance difficulty comparison etc., and summarizes the superiority and the flaw of rule engine framework, also when should use rule engine framework and when should not.