Data Flow Based Integration Testing for Embedded System Using Interaction Model

Abstract

Embedded system covers a large range of highly sensitive area where a slightest error can damage large number of resources even cost human life. So testing embedded system is extremely essential concern in computer science industry. Errors can lies between interacting software and hardware module of embedded system. This interactions can be modeled in terms of data flows between subsystems, i.e. Module, component etc. Tracking data flow across different components is important because applications use this to activate interactions that acquire different values and resources in ways that may ultimately exercise faults. Integration testing is a verification technique of the interfaces among system parts. The integration test of software components is especially concerned with the detection and uncover of interface errors, which are not detectable during unit testing. Existing integration testing approaches typically focused on direct interaction of subsystems which is done by function call but there exists other interactions for indirect dependencies between subsystems, which may occur faults in the system. There are no specific general models for representing all interactions. In this paper we propose an interaction model to cover all possible direct and indirect interaction and provide an integration testing approach using data flow analysis to generate minimal test cases efficiently.