Techniques for auditing and controlling network services
View Patent ↗Techniques for auditing and controlling network services are provided. A proxy is interposed between a principal and a network service. Interactions between the principal and the service pass through the proxy. The proxy selectively raises events and evaluates policy based on the interactions for purposes of auditing and controlling the network service.
1. A method implemented in a non-transitory machine-readable storage medium and processed by one or more processors configured to perform the method, comprising:
intercepting a command instruction sere from a principal in a first processing environment to a network service in a second processing environment;
generating a causation audit event based on the command instruction;
evaluating a policy based on the causation audit event and permitting the command instruction to be passed to the network service for processing in response to a successful evaluation of the policy;
detecting a response from the fretwork service being directed to the principal;
generating an effect audit event;
evaluating a second policy based on the effect audit event and permitting the response to be passed to the principal in response to a successful evaluation of the second policy;
wherein generating the causation event further includes setting a timer that is started when the command instruction is passed to the network service and that is terminated when the response is detected from the network service and recording a relationship within a metrics data store that records a relation for an elapsed time associated with the timer and a state of the second processing environment along with the command instruction, a principal identity for the principal, and a network service identity for the network service; and
capturing a new state of the second processing environment and retrieving the relation for evaluation with the policy when a second instance of the command instruction is received from the principal or a different principal.
2. The method of claim 1 further comprising, using historical metrics to assist in evaluating the policy and the second policy, the historical metrics including response performance values for the network service given particular states of the second processing environment.
3. The method of claim 1 , wherein evaluating the policy further includes interacting with an audit system that is external to the first processing environment to communicate the causation audit event, wherein the policy evaluation dictates that a confirmation is needed from the audit system before concluding successful evaluation of the policy.
4. The method of claim 1 , wherein evaluating the policy further includes using cause and effect profiles from previous audited command instructions associated with the network service to determine that the policy evaluation is successful.