Searching for the Optimal Proof Schedule in a Proof-Carrying Survivability Paradigm - A Dynamic, Interactive Approach

Survivability represents a system's capability to withstand malicious attacks and system failures and to provide essential services in a challenging environment. In a proof-carrying survivability scenario, a user publishes his/her survivability requirements and a system provider constructs a proof to show that the system satisfies those requirements. Finally, the user verifies that the proof is valid. In this paper, we propose a systematic approach for the system provider to search for an optimal proof schedule to construct a valid proof. Our framework is applied to a threshold-based survivability requirement model, where the user's requirements are represented using pre-defined threshold operators and proof generation relies on the evaluations of system survivability properties by authorized evaluators. We develop algorithms and techniques to explore different proof options and schedules and identify the most efficient (optimal) schedule. We study how the dynamic, interactive proof evidence collection process proceeds between the system provider and the evaluators. A prototyping system has been developed to implement the framework.