No abstract

No abstract

Nowadays cloud architecture is widely applied on the internet. New malware aiming at the privacy data stealing or crypto currency mining is threatening the security of cloud platforms. In view of the problems with existing application behavior monitoring methods such as coarse-grained analysis, high performance overhead and lack of applicability, this paper proposes a new fine-grained binary program monitoring and analysis method based on multiple system level components, which is used to detect the possible privacy leakage of applications installed on cloud platforms. It can be used online in cloud platform environments for fine-grained automated analysis of target programs, ensuring the stability and continuity of program execution. We combine the external interception and internal instrumentation and design a variety of optimization schemes to further reduce the impact of fine-grained analysis on the performance of target programs, enabling it to be employed in actual environments. The experimental results show that the proposed method is feasible and can achieve the acceptable analysis performance while consuming a small amount of system resources. The optimization schemes can go beyond traditional dynamic instrumentation methods with better analytical performance and can be more applicable to online analysis on cloud platforms.

No abstract

No abstract