Security policy compliance with violation management

Abstract

A security policy of an information system is a set of security requirements that correspond to permissions, prohibitions and obligations to execute some actions when some contextual conditions are satisfied. Traditional approaches consider that the information system enforces its associated security policy if and only if actions executed in this system are permitted by the policy (if the policy is closed) or not prohibited (if the policy is open) and every obligatory actions are actually executed in the system (no violation of obligations). In this paper, we investigate a more sophisticated approach in which an information system specification is compliant with its security policy even though some security requirements may be violated. Our proposal is to consider that this is acceptable when the security policy specifies additional requirements that apply in case of violation of other security requirements. In this case, we formally define conditions to be satisfied by an information system to comply with its security policy. We then present a proof-based approach to check if these conditions are enforced.

Aiding Compliance Governance in Service-Based Business Processes

“…For instance, in Liu et al 2007 the problem of static (i.e., before process execution) compliance checking of process models against compliance rules is addressed by expressing the models in pi-calculus and the corresponding rules in linear temporal logic; then, model checking techniques are used to determine whether a process model complies with the rules or not. In Brunel et al 2007, policies are modeled and checked as deontic sentences (i.e., rules are of the form "it is obligatory that X..." or "it is permitted that Y..."); then, a system can be compliant even if violations occur, in which case, a second-level set of rules might be applied, for which, again, compliance needs to be checked. A similar modeling technique is presented in Saqid et al 2007 from it and used to annotate the process model so that control concerns can be visualized in the process model space.…”

Usage Control Enforcement - A Survey

“…Subsequently the consumer must provide evidence that the provisional actions have been taken and that he has committed himself to the obligations and compensations. Compensations, or sanctions as they are termed, is also proposed formally in [5]. The authors show how a system may violate certain rules in a policy and still be compliant with the policy through fulfilling a set of sanctions.…”

xESB: An Enterprise Service Bus for Access and Usage Control Policy Enforcement

“…Unaware of any implementation of a generic usage control language that supports compensations, we have developed a proof-of-concept policy language fit for the ESB. Compared to theoretic works on access control compensations [17], violation management [18] and obligation assessment [19,20], we go beyond access control and provide an implementation of compensations on the fly. This means that whenever a violation of some service usage rule is detected, the correction happens as the event travels through the system, before it reaches some interested party.…”