查看文章

As threats to computer security become more common, complex and frequent, systems that can automatically protect themselves from attacks are imminently needed. In this paper, we propose a formal approach to achieve self-protection by performing security analysis on self-adaptive systems, taking the adaptation process into account. We use probabilistic model checking to quantitatively analyze adaptation security, rank the strategies available and select the most secure one to apply in the system. We have incorporated our approach in Rainbow which is a framework to develop architecture-based self-adaptive systems. To evaluate our approach’s effectiveness, we applied it on two case studies: a simple document storage system and ZNN, a well known self-adaptive exemplar. The results show that applying our approach can guarantee a reasonable degree of security, both during and after adaptation.