Study on Integration Test Order Generation Algorithm for SOA

Abstract

Integration test order generation is an important problem in software integration testing research.Reasonable test order can improve the efficiency of integration test and reduce the cost of test.Service oriented architecture(SOA) is a kind of distributed architecture widely used in enterprises in recent years.At present,there are few researches on integration test order generation in SOA architecture.Due to the polymorphism of service composition in SOA architecture,it is impossible to get the integration test sequence between service software in SOA architecture by using the traditional top-down and bottom-up integration test strategies.However,the current research on the generation of integration test sequence based on class cluster in object-oriented system is difficult to apply to the complex coupling relationship between services in SOA architecture.Therefore,an integration test order generation method based on genetic algorithm is proposed to solve the integration test problem between service software in SOA architecture.In this method,the concept of service feature group is used to represent the influence factors of integration testing,and the concept of integration testing priority is used to represent the importance of integration testing of service software.At the same time,the test dependency graph is constructed to describe the complex coupling relationship between ser-vice software in SOA Architecture.In order to reduce the cost of test,a genetic algorithm is designed to generate integrated test sequence.Finally,an example is given to verify the feasibility and correctness of the method.The example results show that the proposed method can generate service software integration test orders with relatively high test priority and low test cost.

Keywords

• soa|web service|integration test order|test dependency graph|genetic algorithm