Input prediction in a database access control system

Abstract

A local database access control system (LDACS) intelligently determines which database access requests intercepted by a database agent requires analysis by an external security device and which of those requests might be predicted not to require such processing e.g., because they do not contain database object information that needs to be validated against a security policy. Client requests that are predicted not to require such processing are then passed to the database server directly without being held by the agent and delivered externally for policy validation. In this approach, the agent does not send every intercepted request to the security device for evaluation against the one or more security policies. Rather, only those intercepted requests that are predicted to contain database object information are delivered. The security device implements an input prediction scheme to facilitate this process by sending control commands to the agent.

Dependencies are inferred from the claim text.