Abstract
Dynamic Information Flow Tracking (DIFT) forms the foundation of a wide range of security and privacy analysis. However, the performance and scalability of DIFT are impeded by the extensive number of states and data flows within programs, which make the execution of interactive data flow analysis queries cumbersome with traditional methods. In this talk, we identify that DIFT under dependency-based information flow rules can be cast as linear transformations over taint states. This transformation enables a novel matrix-based representation, FlowMatrix, to represent DIFT operations concisely and makes it practical to adopt GPUs as co-processors for DIFT analysis. FlowMatrix provides efficient support for interactive DIFT query operations. We design a DIFT query system and prototype it on commodity GPUs. Our evaluation shows that our prototype outperforms CPU-based baseline by several times and enables rapid response to DIFT queries. It has two to three orders of magnitude higher throughput compared to typical DIFT analysis solutions. FlowMatrix represents an initial step in the direction toward balancing the complexity and efficiency to adopt new processing technologies in information-flow analysis. We also discuss the potential and limitation of related technologies through the view of complexity.
Lecturer
Zhenkai Liang is an Associate Professor in the Department of Computer Science at National University of Singapore. He is also a co-Lead Principal Investigator of National Security R&D Lab of Singapore. His research interests are in system and software security, such as binary program analysis, security in Web, mobile, and Internet-of-things (IoT) platforms. He has been publishing high-impact papers in top security and software engineering conferences, and has won several best paper awards in security and software engineering conference, including Annual Computer Security Applications Conference (ACSAC), USENIX Security Symposium, and ACM SIGSOFT Symposium on the Foundations of Software Engineering (FSE). He has also won the Annual Teaching Excellence Award of NUS in 2014 and 2015. He is a current member of the Steering Group of NDSS and has served as technical committee members and editorial board members of main security conferences and journals, including ACM Conference on Computer and Communications Security (CCS), USENIX Security Symposium, Network and Distributed System Security Symposium (NDSS), and IEEE Transactions on Dependable and Secure Computing (TDSC) and ACM Transaction on Privacy and Security (TOPS). He received his Ph.D. degree in Computer Science from Stony Brook University in 2006, and B.S. degrees in Computer Science and Economics from Peking University in 1999.
