CDCAS: A Novel Cloud Data Center Security Auditing System

With the increasing acceptance of cloud data center and virtualization technology by enterprises and industries, the security concern becomes the key hindrance to the development and deployment of cloud computing. Security auditing is a good way to deal with the threats faced by a cloud data center. But traditional auditing is no longer suitable for the new cloud environment. In this paper, we design, implement and evaluate the CDCAS, a novel cloud data center auditing system, which matches the demand of the scalability and efficiency of a cloud data center. In this system, we design one distributed and autonomous agent model which can be controlled by a set of rules dynamically generated to fit its use scenario. We then build the log analysis model which uses the signature based method and correlative analysis algorithm to extract security events from collected log with agreeable false positives. We evaluate our system both on real world and simulation to validate its efficiency. And our system is also deployed by the cloud data center of a well-known financial institution, and performs well.