RINGA: Design and verification of finite state machine for self-adaptive software at runtime

• Published in: Information and software technology 2018-01, Vol.93, p.200-222
• Main Authors: Lee, Euijong, Kim, Young-Gab, Seo, Young-Duk, Seol, Kwangsoo, Baik, Doo-Kwon
• Format: Article
• Language: English
• Subjects: Finite state machine; Model checking; Runtime; Self-adaptive software
• ISSN: 0950-5849; 1873-6025
• DOI: 10.1016/j.infsof.2017.09.008

•A self-adaptive framework for runtime verification using model-checking is proposed.•A finite-state model for designing self-adaptive software is proposed.•Performance evaluation of the self-adaptive framework verification is conducted.•Applying the self-adaptive framework into IoT based scenario. In recent years, software environments such as the cloud and Internet of Things (IoT) have become increasingly sophisticated, and as a result, development of adaptable software has become very important. Self-adaptive software is appropriate for today's needs because it changes its behavior or structure in response to a changing environment at runtime. To adapt to changing environments, runtime verification is an important requirement, and research that integrates traditional verification with self-adaptive software is in high demand. Model checking is an effective static verification method for software, but existing problems at runtime remain unresolved. In this paper, we propose a self-adaptive software framework that applies model checking to software to enable verification at runtime. The proposed framework consists of two parts: the design of self-adaptive software using a finite state machine and the adaptation of the software during runtime. For the first part, we propose two finite state machines for self-adaptive software called the self-adaptive finite state machine (SA-FSM) and abstracted finite state machine (A-FSM). For the runtime verification part, a self-adaptation process based on a MAPE (monitoring, analyzing, planning, and executing) loop is implemented. We performed an empirical evaluation with several model-checking tools (i.e., NuSMV and CadenceSMV), and the results show that the proposed method is more efficient at runtime. We also investigated a simple example application in six scenarios related to the IoT environment. We implemented Android and Arduino applications, and the results show the practical usability of the proposed self-adaptive framework at runtime. We proposed a framework for integrating model checking with a self-adaptive software lifecycle. The results of our experiments showed that the proposed framework can achieve verify self-adaptation software at runtime.