Given the many advantages offered by cloud computing, organizations are increasingly moving their databases to cloud. Some of these databases contain critical infrastructure data requiring real-time service. Unfortunately, data in critical infrastructure systems has become one of the major targets of attackers. Since such data are highly connected and interdependent, the initial damage done by an attacker spreads quickly through the system when valid transactions make any updates based on the value of a damaged object. As a result, it adversely affects the real-time service the system is designed to offer. In this work, we provide a model to perform fast damage assessment and recovery. Our approach allows all unaffected data objects to be available to users and makes the affected system recover swiftly. This work uses a modified log mechanism and a graph-based approach to perform fast damage assessment and recovery. Through simulation, we proved that the model expedites the process significantly.