The architecture of an automatic distributed recovery system

Distributed computing systems are important in daily business processing. Since vulnerability cannot be totally removed from a distributed computing system, successful attacks always happen and may inject malicious tasks or incorrect data into the distributed system. Referring to the incorrect data will further corrupt more data objects in the system, which comprises the integrity level of the system. In this paper, based on our recovery theories and algorithms proposed in [Yu, M., et al. (2004)], we build an architecture of a distributed attack recovery system. Our work can help the user to build a real recovery system and serve as a test bed to evaluate both our attack recovery techniques and new developed attack recovery techniques in the future.