The Community for Technology Leaders
Autonomic Computing, International Conference on (2005)
Seattle, Washington
June 13, 2005 to June 16, 2005
ISBN: 0-7965-2276-9
pp: 263-273
Mrinmoy Ghosh , Georgia Institute of Technology
Guofei Gu , Georgia Institute of Technology
Weidong Shi , Georgia Institute of Technology
Trevor N. Mudge , University of Michigan, Ann Arbor
Hsien-Hsin S. Lee , Georgia Institute of Technology
Laura Falk , University of Michigan, Ann Arbor
This paper proposes a novel system design using a chip multiprocessor (CMP) to provide intrusion tolerance and self-recovery for server applications. Our platform provides three major advantages over previously proposed approaches, 1) security insulation from remote exploits and attacks; 2) close coupling between processor cores in a CMP to ensure immediate logging, fine-grained inspection and fast recovery; 3) concurrent and fine-grained inspection, logging and recovery techniques that are off of the critical path. We have designed a multi-point defense and recovery system to defeat remote exploits. We used a checkpoint based approach to recover server applications under attack. It takes a snapshot of the application?s context and memory state before it handles the next request. If the request turns out to be malicious, the system can discard the malicious request and rollback the application?s state to a known good one through checkpointing. We have also designed an rapid recovery system for kernel space rootkit attacks. Our intrusion survivable and self-recovery design provides reliable production services that System Administrators are seeking.
Intrusion-tolerant computing, survivable service, chip multi processor, self-healing, rootkits, buffer overflow.
Mrinmoy Ghosh, Guofei Gu, Weidong Shi, Trevor N. Mudge, Hsien-Hsin S. Lee, Laura Falk, "An Intrusion-Tolerant and Self-Recoverable Network Service System Using A Security Enhanced Chip Multiprocessor", Autonomic Computing, International Conference on, vol. 00, no. , pp. 263-273, 2005, doi:10.1109/ICAC.2005.8
102 ms
(Ver 3.3 (11022016))