Modeling and reliability evaluation of Avionics Clouds based on AADL and GSPN

The proposed concept of IMA2G or DIMA supports increased computing performance, provides abstraction of platform level services, and implements reconfiguration mechanisms. With improving performances, the architecture is becoming more complex, which increases the cost. Considering cloud computing applied in commercial world achieves a great success in providing more services with lower costs, the concept of Avionics Clouds is put forward, which integrate all the resources of platforms from air, space and ground to provide services for completing missions of aircrafts. In order to meet the criticality of avionics applications, it is necessary to evaluate the reliability of Avionics Clouds quantitatively. In this paper, a study case of the Avionics Clouds is first proposed, and it is modeled using Architecture Analysis and Description Language (AADL). With specified error model injected in, the AADL error model is then converted into Generalized Stochastic Petri Net (GSPN) model according to a series of transformation rules. At last, the GSPN model is analyzed by PIPE2, and reliability of the case is calculated by the possibility of stable states. The results show that reliability of Avionics Clouds could satisfy the requirements of avionics applications.