Trusted Recovery from Information Attacks

Abstract : Preventive measures sometimes fail to deflect malicious attacks. In this work, we adopt an information warfare perspective which assumes success by the attacker in achieving partial, but not complete damage. in particular, we work in the database context and consider recovery form malicious but committed transactions. Traditional recovery mechanisms do not address this problem, except for complete rollbacks, which undo the work of benign transactions as well as malicious ones, and compensating transactions, whose utility depends on application semantics. recovery is complicated by the presence of benign transactions that depend, directly or indirectly, on the malicious transactions. We present recovery models to restore only the damaged part of the database. Two families of new repair algorithms are developed: one is a set of dependency-graph based algorithms, the other is a set of algorithms that do repair via rewriting histories.