To date implementations of Internet of Things (IoT) architectures are confined to particular application areas and tailored to meet only the limited requirements of their narrow applications. To overcome technology and sector boundaries this paper proposes a dynamic service creation environment that employs i) orchestration of business services based on re-usable IoT service components, ii) self-management capable components for automated configuration and testing of services for things, and iii) abstraction of the heterogeneity of underlying technologies to ensure interoperability. To ensure reliability and robustness the presented approach integrates self-testing and self-adaptation in all service life cycle phases. The service life cycle management distinguishes the IoT service creation phase (design-time) and the IoT service provision phase (run-time). For test-friendly service creation (1) semantic service descriptions are employed to derive semi-automatically services and related tests, (2) and testing is systematically integrated into a Service Creation Environment. For reliable and robust service provisioning the presented system (3) forces validation tests in a sandbox environment before deployment and (4) enables run-time monitoring for service adaptation. The system under test is modelled by finite state machines (FSM) that are semi-automatically composed of re-usable test components. Then path searching algorithms are applied to derive automatically tests from the FSM model. The resulting tests are specified in the test control notation TTCN-3 and compiled to run the validation tests.
展开▼
