It is particularly challenging to defend common distributed systems against security vulnerabilities because of the complexity and their large sizes. However, traditional solutions, that attack the information flow security problem, often fail for large, complex real-world distributed systems due to scalability problems. The problem would be even exacerbated for the online defense of continuously-running systems. My proposed research consists of three connected themes. First, I have developed metrics to help users understand and analyze the security characteristics of distributed systems at runtime in relation to their coupling measures. Then, I have also developed a highly scalable, cost-effective dynamic information flow analysis approach for distributed systems. It can detect implicit dependencies and find real security vulnerabilities in industrial distributed systems with practical portability and scalability. In order to thoroughly solve the scalability problem in general scenarios, I am developing a self-adaptive dynamic dependency analysis framework to monitor security issues during continuous running. In this proposal, I outline the three projects in a related manner as to how they consistently target the central objective of my thesis research.

@inproceedings{paperbot5369,
  title = {Towards scalable defense of information flow security for distributed systems},
  author = {Xiaoqin Fu},
  booktitle = {ISSTA},
  year = {2019},
  doi = {10.1145/3293882.3338988}
}