Towards Achieving Data Security with the Cloud Computing Adoption Framework

Offering real-time data security for petabytes of data is important for Cloud Computing. A recent survey on cloud security states that the security of users’ data has the highest priority as well as concern. We believe this can only be able to achieve with an approach that is systematic, adoptable and well-structured. Therefore, this paper has developed a framework known as Cloud Computing Adoption Framework (CCAF) which has been customized for securing cloud data. This paper explains the overview, rationale and components in the CCAF to protect data security. CCAF is illustrated by the system design based on the requirements and the implementation demonstrated by the CCAF multi-layered security. Since our Data Center has 10 petabytes of data, there is a huge task to provide real-time protection and quarantine. We use Business Process Modeling Notation (BPMN) to simulate how data is in use. The use of BPMN simulation allows us to evaluate the chosen security performances before actual implementation. Results show that the time to take control of security breach can take between 50 and 125 hours. This means that additional security is required to ensure all data is well-protected in the crucial 125 hours. This paper has also demonstrated that CCAF multi-layered security can protect data in real-time and it has three layers of security: 1) firewall and access control; 2) identity management and intrusion prevention and 3) convergent encryption. To validate CCAF, this paper has undertaken two sets of ethical-hacking experiments involved with penetration testing with 10,000 trojans and viruses. The CCAF multi-layered security can block 9,919 viruses and trojans which can be destroyed in seconds and the remaining ones can be quarantined or isolated. The experiments show although the percentage of blocking can decrease for continuous injection of viruses and trojans, 97.43% of them can be quarantined. Our CCAF multi-layered security has an average of 20% better p- rformance than the single-layered approach which could only block 7,438 viruses and trojans. CCAF can be more effective when combined with BPMN simulation to evaluate security process and penetrating testing results.

http://dx.doi.org/10.1109/TSC.2015.2491281